Latest Updates





South East spring show 2015 icon
Click for details



Anoxic filtration03               Anoxic filtration


This article has been written so as to be as easy to read as possible for those who would like the simplest possible explanation of the very complex subject of anoxic filtration. It also contains science panels that give more detailed information for those who might like to know how it works in greater detail. (For simplicity, the baskets used in the anoxic filtration system will be referred to throughout this article as “baskets” or “biocenosis baskets” but their proper name is Biocenosis Clarification Baskets”).

Anoxic Filtration – is it a bog filter?




Kitty litter




Take a planting basket, fill it with cat litter, then scoop out a depression from the centre and fill that depression with about 400 grams of an aquarium plant fertiliser called laterite. Put a plant into the basket if you wish, but it isn’t necessary.

If it’s that simple to make a biocenosis basket, why are more koi keepers not using the anoxic filtration system?

Possibly the answer lies in the fact that it’s rare to find a koi keeper who really understands what anoxic filtration actually is, what it does and how it works. The reason for this lack of understanding should immediately be obvious.

Already, I have used three words that many koi keepers will find new and mysterious. What is laterite? What is a biocenosis basket? And isn’t anoxic, something to do with anaerobic sludge where harmful bacteria can live? If it were possible to describe the anoxic system without using such words I would do so, but, whilst it’s easy to build the system, the way it works is extremely complicated. So let us begin by explaining what laterite, biocenosis and anoxic mean.

Laterite: This is easy enough to understand. It’s simply a clay based material that is rich in iron. It is used in tropical fish tanks as a plant fertiliser and can be brought on-line or from tropical fish dealers.

Biocenosis: This is a scientific term for a place where different biological processes take place, each process being of mutual benefit to the other processes.

Anoxic: Anoxic does not mean the same as anaerobic. In an anaerobic environment there is a complete absence of oxygen. In an anoxic environment, there is oxygen present but it is at a very low level. In a biocenosis basket there is always a low level of oxygen. Levels are typically between 0.5 mg/L and 2 mg/L.

This is the key factor that will influence a situation where anoxic filtration can occur. The presence of an extremely low level of oxygen is crucial to the system as will be described later.

....the anoxic filter system is quite unlike any filter system that is commonly in use by pond keepers....

In passing, it might be worthwhile to contrast the oxygen level in a biocenosis basket with the oxygen level in the pond itself. The minimum acceptable oxygen level in a koi pond is 6 mg/L, although 7 mg/L is more often recommended as a safer minimum to adopt and, in practice, it should be as near saturation point as is possible. (Saturation point is the highest level of oxygen that will dissolve into water at any given temperature).

With the first few terms explained, it should now be possible to move on to a better understanding of how the anoxic filter system is quite unlike any filter system that is commonly in use by pond keepers. To contrast anoxic filtration with conventional filtration, it is first necessary to understand how biological filters actually work.

Fish ammonia excretionsConventional biological filtration
Fish continuously excrete ammonia which is toxic to them, and so it has to be removed from the pond water before it can cause them harm.

Any well designed conventional biological filter system will be effective at taking this ammonia and converting it, first into nitrite, and then into nitrate by a process called the “nitrogen cycle”.

Conventional (nitrogen cycle) filtration very roughly takes one ammonia molecule, four oxygen molecules and seven carbonate molecules into the cycle, and after the bugs have done their job, there is one nitrate molecule excreted as a waste product. Between the two main species of bugs we could say that, in round numbers, one ammonia molecule becomes one nitrate molecule and consumes seven carbonates and four oxygens to do so.

Having achieved the conversion of ammonia into nitrate, the task of a conventional biological filter is complete. No further biological action to remove the nitrate takes place and so the level of nitrate in the water slowly rises. This is the first disadvantage of conventional biofilter systems.

Nitrate is nowhere near as toxic to koi as either ammonia or nitrite but that doesn’t mean that they are not affected by it. Hard scientific research on this subject is difficult to come by, but there is plenty of anecdotal evidence to show that koi kept in low nitrate ponds, show better growth and colour development than those kept in a pond where the nitrate level has been allowed to rise.

Science panel

Nitrate affecting shiroji
 Takayuki Izeki, on behalf of the ZNA, conducted some research into the effect of nitrate on the shiroji (white) of kohaku.
 He found that the maximum acceptable level of nitrate for koi keeping should be 15 mg/L, saying: "anything more than this and the skin gradually begins to deteriorate but will improve again if the concentration is reduced."
 Takayuki-san also said that if the pond nitrate level could be kept lower than 5 mg/L, "the skin becomes so white it virtually shines".
 His research showed that the negative effect of nitrate was caused by it reducing the koi’s ability to discharge metabolic toxins from its body and that this resulted in these metabolic wastes building up in the skin.

Easier to prove, is the effect that nitrate has on algae of all types. Nitrate is a plant fertiliser. A rising level will encourage filamentous algae (cyanobacteria or blanket weed) and planktonic algae (floating algae that turns water green). Water changes will help but, even after a 10% water change, the nitrate level will still be 90% of the original value, and, as more ammonia is converted to nitrate, the level will soon begin rising again.

A second problem with some conventional biological systems is that the media can become anaerobic (no oxygen at all).

This will not occur with moving media such as fluidised sand or aerated K1, but where the media is static, water flowing through it carries suspended particulate matter as it passes through. This will settle within the media and, over time, can cause it to block. Water can’t easily pass through blocked areas and will tend to bypass them because it will find it easier to flow through areas that are not blocked (yet!). This is why K.O.I. does not recommend some of the older filtration media such as lava rock. Water flowing through media carries oxygen to the bacteria that are living within it. As the flow through a particular area reduces, the bacteria in it will find that there is less and less oxygen in their environment.

The nitrifying bacteria that have been oxidising ammonia firstly to nitrite and then to nitrate will have been using a little over four times as much oxygen as the ammonia they are oxidising.  As the media becomes clogged, they will find that they no longer have the oxygen available to carry on this process. There are other bacteria, called facultative anaerobes, that need oxygen as part of their biochemistry, but in some ways, can be thought of as being far more clever than we are. If we are deprived of oxygen, we soon die. These bacteria normally take the oxygen that they need directly from the water surrounding them, but if there is no oxygen in that water, there is a way they can get it.

They can take oxygen from nitrate. Nitrate (NO3) is the end product of the nitrogen cycle, (as far as koi keepers are concerned). It consists of one atom of nitrogen joined to three atoms of oxygen. However, when these bacteria are in biofilter media that has become blocked and therefore poorly oxygenated, and they take away the atoms of oxygen from nitrate, they reverse the nitrogen cycle.

Beneficial filter bacteria will be as busy as ever converting ammonia to nitrate in the aerobic areas of the biological filter, but bugs that can live in anaerobic conditions will rapidly colonise the anaerobic areas and will become equally busy taking the oxygen that they need from the nitrate that has just been produced by their cousins. This will result in that nitrate being converted back to ammonia again before it leaves.

This is a completely pointless exercise, yet it is exactly what is happening in many filter systems where the biological media is not clean. At least part of the good work being done by the aerobic bugs in the oxygen-rich areas, is being undone by anaerobic bugs in areas that are deprived of oxygen.

Understanding anoxic filtration
The anoxic filtration system was designed and has been developed over many years by Dr Kevin Novak PhD and it addresses both these problems. The anoxic system doesn’t rely on converting ammonia to nitrite and then into nitrate.

What is less well known is that plants actually “prefer” ammonia to nitrate as a nutrient source.

With the anoxic system, ammonia is either converted directly to nitrogen gas, (more correctly called dinitrogen (N2) because nitrogen atoms naturally form into pairs) by bacterial action in the unplanted baskets, or it is taken up by plant roots in baskets that also contain plants.

It’s well known that plants “like” nitrate as one of their nutrients. What is less well known is that plants actually “prefer” ammonia to nitrate as a nutrient source, and will take it directly from pond water if it can be presented to their roots and held there. The process that causes this to happen is described in the science panel, but simply put, ammonia molecules are attracted into the baskets by the negative electrical charges that exist within it and that plant roots need to expend less energy taking up ammonia than is needed to take up nitrate.

Science panel

Plants “prefer” ammonia to nitrate
 Plants need nitrogen in order to make amino acids which are then combined together to make their proteins. They can take up nitrogen either as nitrate (NO3) or ammonia (NH3).
 However, using nitrate is more costly to the plant in terms of energy because nitrate has to be converted by the plant first into nitrite and then into ammonia.
 Only after nitrate has been converted into ammonia can it be incorporated into the plant amino acids.
 On the other hand, once ammonia has been taken up by the roots it becomes immediately available to be incorporated into amino acids without the plant first having to expend energy as it does with nitrate.
 In fact nitrate uptake by plants requires so much energy that they can only do it while they are making energy by photosynthesis in the daytime.
 However, taking up their preferred nutrient, ammonia, requires so little energy that they can continue to take it up through the night as well as during the day.

Another important feature to understand is that it’s only the ammonia molecules that are drawn into the baskets. Obviously, water floods into them when they are immersed, but after that, water doesn’t actually need to flow through them in order to filter out ammonia. The electrical charges in the centre of the basket only draw in ammonia molecules; they don’t draw in water molecules.

This directly addresses the second possible problem that can happen with some of the static types of conventional filter media – suspended particulates can clog the media and it will then become anaerobic. Critics, who haven’t taken the trouble to understand how the anoxic system works, often wrongly describe it as a “bog filter, full of nasty anaerobic bacteria”. They warn that the baskets are a breeding ground for parasitic bugs that can then spread to your fish. In fact the direct opposite is true.

Biocenosis baskets cannot clog because, if no water flows through them, there is no way that debris can be carried inside. On the other hand, if water flowing through conventional biofilters doesn’t have every speck of debris filtered out of it, there will always be the risk that sludge will settle inside and block the media. So, far from a biocenosis basket being a “bog filter”, it’s more likely that this label could be applied to a conventional system that hasn’t been kept sufficiently clean!

The anoxic filtration system
Figure 1 shows a typical anoxic filtration pond without the plants. It’s simply a shallow pond, around 60 cm (24 inches) deep, with water being pumped into one end and overflowing by gravity back to the main pond at the other.

Anoxic pond without plants02
Figure 1
A typical anoxic pond without the plants

This one contains 22 biocenosis baskets. The small pebbles on top of the baskets prevent water flow from disturbing the cat litter inside and causing it to float away. Figure 2 shows how the entire anoxic filtration system can be “hidden in plain sight”.

Image48 (250)
Figure 2
The entire filter system can be hidden in plain sight. The top pond is actually the anoxic filter

The anoxic pond looks like a water garden, not a filter, and yet everything except the pump in the main pond can be clearly seen. For those who don’t like pump fed systems, the anoxic pond can be gravity fed from a bottom drain. A modification to the original pump fed design for those who prefer filters that are gravity fed from their bottom drain is shown in figure 5, when the various design layouts are discussed later in this article.

The simplest design is to build the anoxic pond at the same level as a conventional gravity fed system and use a submersible pump or external dry mounted pump to pump water from the anoxic pond back into the main pond. The only limit to how this system can be built or adapted is your ingenuity!

There are few hard and fast rules as to how to make biocenosis baskets. It is important that the planting baskets used should have open lattice type sides to allow ammonia to be drawn in through them, but apart from that, any basket around 30 cm x 30 cm x 20 cm deep will do. The cat litter should be a granular, unscented, “non-clumping” type that should retain its granular structure when wet.

Science panel

Ammonia to nitrate conversion
 It’s easy to remember the nitrogen cycle as approximately converting one molecule of ammonia (NH4) into one molecule of nitrate (NO3).
 For those who want to know the accurate numbers; during the process 220 molecules of NH4 become 216 molecules of NO3.
 Almost a one-to-one conversion but not quite.

The only limit to how this system can be built or adapted is your ingenuity!

Laterite can be purchased on-line in 1.6 kg packets for less than £20 / US$33 / €25. One packet is sufficient for four full size baskets so the total cost of laterite for a 24 basket system will be around £120 / US$200 / €150 [2014 prices and currency conversion].

In recent years there have been more and more hobbyists who have asked if biocenosis baskets can be added to a pond that already has a mature conventional biofilter. The answer is “yes”, although anoxic filtration was designed to replace conventional filtration rather than to be an addition.

However, for those who might like to try the idea first before considering whether to replace their conventional filters, “yes”, it is possible to add anoxic filtration to a conventional filter system. This is because a biocenosis basket, not only removes ammonia without leaving any nitrate, but has sufficient spare capacity to remove nitrate generated by nitrogen cycle biological processes going on elsewhere in the pond or even in an existing conventional biological filter.

Biocenosis baskets need very little maintenance because the life of the cat litter is indefinite and the laterite is only very slowly depleted by plants. Because plants use photosynthesis to make energy from sunlight and a green pigment called chlorophyll is essential for this process. Plants need iron to make chlorophyll so, if there is a plant in the biocenosis basket, the iron in the laterite will become exhausted after about five to ten years and you will have to add some more laterite or remake the baskets

Nitrifying bugs are everywhere
Nitrifying bugs (nitrogen cycle bugs such as nitrosomonas and nitrobacter) are abundant in nature. They aren’t just confined to an aquatic life; they are soil bacteria and are just as happy out of water as in water. In natural ponds, lakes or rivers they grow on every available surface including the pond bottom, rocks and plants. In an ornamental pond they will also grow on all available surfaces, not just within the filter.

Science panel

The simplified biochemistry of the nitrogen cycle

 What nitrosomonas eat for lunch:-
55NH4+ + 76O2 + 109HCO3- = C5H7O2N + 54NO2- + 57H2O + 104H2CO3

 What nitrobacter eat for lunch:-
400NO2- + NH4+ + 4H2CO3 + HCO3- + 195O2 = C5H7O2N + 3H2O + 400NO3-

What the equations tell us
 You can either count atoms and molecules, or you can take my word for it, that these equations could very roughly be described as saying:-
 One molecule of ammonia + four molecules of oxygen + seven molecules of carbonate become one molecule of nitrate + a bit of bug tissue (the C5H7O2N in these equations is a molecule of “bug”). In other words, the bacteria can be thought of as consuming ammonia, oxygen and carbonate and getting slightly bigger. (Eventually when they have consumed enough, each bug will divide into two separate bugs, but that is a topic for another article).
 If we ignore the carbonate and also ignore the fact that the bugs are getting bigger in this process and we just concern ourselves with what happens to the ammonia, it gets even simpler.
 One ammonia molecule + four oxygen molecules (plus seven carbonate and four oxygen molecules) become one nitrate molecule.
 The point to note from these equations is that nitrifying bacteria cannot convert ammonia to nitrate without a plentiful supply of oxygen (and carbonates).

The purpose of a biological filter is to make a “bacteria friendly” environment that will concentrate the bulk of the population in one easy to manage area where the main nitrogen cycle will occur. However, ammonia is also being nitrified throughout the whole pond. All that is necessary for this to occur is a wet surface and a supply of ammonia, oxygen and carbonate.

Taking all the media out of the biofilter in a conventional filtration system, will not stop the production of nitrate altogether, it will still be produced elsewhere.

It may not be immediately obvious but there are also ample opportunities for the nitrogen cycle to take place within the baskets themselves.

The baskets are under water so, stating the obvious, all surfaces of the clay particles are wet. The water just inside the baskets will also be rich in oxygen and carbonate, so it is an ideal place for nitrifying bugs to set up home and to convert ammonia firstly to nitrite and then to nitrate.

If nitrate is actually produced within a basket that is designed to eliminate nitrate, does this mean that these baskets are a failure? Not in the least, as will be described next.

When nitrifying bugs convert ammonia to nitrate the total amount of oxygen that they use in the process is a little more than four times greater than the amount of ammonia that they are converting!

There are two equations that define what goes on as filter bugs convert ammonia through nitrite and into nitrate. I light-heartedly refer to them as; “what nitrosomonas and nitrobacter eat for lunch”. They have been simplified as far as is possible and are included in the science panel for those that may be interested.

For those who may have skipped the panel, there is one important point that should be noted. When nitrifying bugs (nitrosomonas and nitrobacter) convert ammonia, through nitrite and into nitrate, the total amount of oxygen that they use in the process is a little more than four times greater than the amount of ammonia that they are converting!

This is why we should ensure that conventional biofilters are well aerated because this use of oxygen isn’t optional, it’s essential. Unless that amount of oxygen is available to the aerobic nitrifying bugs that we are familiar with, they cannot process ammonia and the nitrogen cycle comes to a dead stop. This is what happens inside a biocenosis basket; as soon as the dissolved oxygen level in it falls below about 2 mg/L, nitrifying bacteria cannot function and stop converting ammonia into nitrate.

Biocenosis basket
Figure 3
For clarity, zones A and B have not been drawn to scale. In practice these two zones are only a few millimetres thick before nearly all the oxygen has been exhausted

Inside a biocenosis basket
Let us now apply this to what is going on inside a biocenosis basket, and follow the ammonia molecules as they are drawn into the baskets to their doom. See figure 3.

Zone A
Negative charges in the cat litter in the baskets begin to attract molecules of ammonia (NH4) towards its centre. As these molecules pass through zone A nitrifying bacteria (nitrogen cycle bugs) will grab one ammonia molecule and four oxygen molecules; they will then excrete one nitrate molecule.

The ammonia level in zone A will have dropped a little and the nitrate level will have risen by roughly the same amount but the oxygen level will have dropped considerably, (just over four times as much as the reduction in the ammonia level).

Science panel

Facultative anaerobic bacteria
 These are bacteria that are able to live and thrive in aerobic (oxygen rich) conditions but which are also able to function in an anaerobic environment (depleted oxygen - typically between 0.5 mg/L and 2 mg/L).
 All living creatures, including bacteria, need to make a compound called ATP (adenosine triphosphate). ATP is a high energy molecule that has earned the nickname “the energy currency of life” and which is essential for the normal functioning of all living cells.
 When oxygen is present, these bacteria will use it as what is known as a terminal electron acceptor in order to make ATP. Where oxygen is below about 2 mg/L they switch to using the oxygen stolen from other compounds such as nitrate or phosphate as an electron acceptor.
 In that process, nitrate (NO3) or phosphate (PO4) are reduced either to dissolved nitrogen gas, which bubbles away to the atmosphere, or dissolved phosphorous which is also metabolised by the bacteria as it is also required in the production of ATP.

There is now far less oxygen but there will still be enough for the nitrogen cycle to continue so we will continue to journey with the ammonia molecules into zone B.

Zone B
As more and more ammonia is converted to nitrate, the ammonia level drops even more and the nitrate level rises. So much oxygen has been used in the process that this area can no longer be called truly aerobic, (oxygen rich), but there is still a little oxygen left to sustain some nitrifying bacteria so we will follow the remaining ammonia as it journeys into zone C.

Zone C
The ammonia is still being pulled remorselessly toward the centre of the basket but almost all oxygen in the water has already been used. The nitrogen cycle, as we know it, ceases. Nitrification cannot occur if dissolved oxygen levels fall too far below about 2 mg/L and it will be lower than that in zone C where facultative anaerobic heterotrophic bacteria live. So what on earth is a facultative anaerobic heterotrophic bug anyway?

Molecule - nitrate
A molecule of nitrate (NO3) consists of one nitrogen atom (N) joined to three oxygen atoms (O)

“Facultative anaerobic” means that a bug has the ability to live anaerobically, (without oxygen), provided it can steal some. “Heterotrophic” means that it obtains its nutrients from organic molecules.

So a facultative anaerobic heterotrophic bug is simply a bug that can live where there is very little oxygen and gets its nutrients from organic molecules. That wasn’t so hard, was it?

It isn’t hard to understand where a facultative bug can steal its supply of oxygen from either. Remember the nitrate that is being produced by the nitrogen cycle bugs? The chemical symbol for nitrate is NO3, (one atom of nitrogen, joined to three atoms of oxygen). A facultative anaerobic heterotroph can easily make use of the three oxygen atoms and leave just the nitrogen atom which is initially dissolved in the water but will migrate out of the basket and gas off to the atmosphere.

Bugs don’t actually eat
Although it’s convenient to refer to nitrifying bugs as “eating” ammonia or nitrite and breathing oxygen, in reality, they don’t have little mouths nor indeed do they have lungs. Oxygen, and ammonia or nitrate, and other nutrients are simply absorbed directly through their cell walls, just as if we were able to eat by placing food onto our stomachs or breathe by absorbing oxygen through our chests.

So although there are nitrogen cycle bugs living in zones A and B in the biocenosis baskets and they will be busy putting nitrate into the water, other bugs in that same basket are just as busy disposing of it. The overall effect of a basket is to totally remove ammonia with no nitrate or any other by-product chemicals remaining in the water.

Molecules are not little magnets, but for a basic understanding of how molecules work, it’s convenient to imagine that they behave just like little magnets

There’s more
If that was all a biocenosis basket achieved, it would be pretty marvellous, but there is even more science going on. We haven’t even considered the full extent of what is going on deep in the baskets yet, other than to say that “electrical charges” attract ammonia molecules toward the centre of the basket. How does it do this, and what happens to the ammonia when it gets there?

Molecules are not little magnets, but for a basic understanding of how molecules work, it’s convenient to imagine that they behave just like little magnets. When we played with magnets as children, we discovered that two similar magnetic poles repel each other but opposite poles attract and will stick together. Molecules behave just like that but the forces are electrical charges similar to static electricity not magnetism.

Science panel

Facultative anaerobic respiration
Facultative anaerobic bacteria can use normal aerobic respiration when oxygen is present or, where the oxygen level is too low, they can switch to a mode of respiration called fermentation.
 There are different forms of fermentation that facultative anaerobes can use and reducing nitrate to nitrogen gas is one example that is useful to koi keepers but this requires an anoxic environment, since where sufficient oxygen is available, anaerobic respiration is generally shut down and these bacteria respire aerobically.

The charge on an ammonia molecule (NH4+) is positive, and the charges in the basket are negative. The strength of these charges increases towards the centre of the basket. Opposite charges attract and the greater the opposite charge, the greater will be the strength of the attraction and so ammonia molecules will be pulled inside through zones A and B into zone C as described above.

Although some of the ammonia will have been totally disposed of along the way, much will still remain, and once it is there, it cannot escape. The way ammonia is taken up by plants roots is a complex relationship involving yet more molecular charges and it isn’t necessary to understand this mechanism in order to understand how biocenosis baskets work. It’s sufficient to say that the charges attract ammonia right up to the plant roots where it will be absorbed.

What happens in unplanted baskets? Good news – more bugs. For those biocenosis baskets that don’t contain plants, the facultative anaerobic bacteria that inhabit zone C will perform a second clever trick.

Earlier, we discovered that these bacteria preferred to take oxygen directly from the pond water, but when there was little or no oxygen available, as in zone C, their first trick was to obtain some by taking the atoms of oxygen from any nitrate that had been produced by the nitrifying bacteria in zones A and B. When the bugs have used all of the available nitrate, their second trick is to switch to directly metabolising ammonia to provide their energy needs! The expression “clever as a sack of monkeys” should be changed to “clever as a basket of bugs.”

Planted biocenosis baskets are more efficient but whether or not the baskets contain plants, the ammonia that is drawn into a basket has no escape. If plants don’t get it, the bugs will. The fact that the biocenosis baskets don’t have to contain plants to mop up ammonia because a colony of bugs will soon develop and will take the opportunity of a free ammonia lunch, enables anoxic filtration to be sited indoors or disguised under decking.

When the bugs have used all the available nitrate they switch to directly metabolising ammonia to provide their energy needs...
....whether or not the biocenosis baskets contain plants, the ammonia that is drawn in has no escape. If plants don’t get it, the bugs will.

How much nitrate does one basket remove?
Some hobbyists are keen to try anoxic filtration but reluctant to replace an existing matured biofilter with anoxic filtration and ask whether adding a few biocenosis baskets to their existing biofilter will reduce their nitrate level.

It’s impossible to accurately predict just how much a single basket or a small group of baskets will reduce the nitrate level in a pond that has an existing conventional biofilter because every situation will be different. However, we can make an estimate by taking into account the recommendation that, for a pond that is biologically filtered just by anoxic filtration, there should be about one biocenosis basket per adult fish. If fewer baskets than that number is added to a pond with existing biofiltration what would be the effect?

Science panel

Estimating nitrate reduction from a few baskets
 If a pond was filtered by two identical conventional filters working in parallel (side by side) it could be expected that each filter set would take about half of the available ammonia and be responsible for producing about half of the resulting nitrate.
 If the two conventional filters are not equal in capability, the ammonia will be converted to nitrate by each filter in approximate proportion to its size or capability.
 If one of the filters was a full size filter and the other filter was smaller that only had sufficient capability to handle, say, 25% of the ammonia then, very approximately, we could estimate that the full size filter will take 80% of the ammonia (its 100% capability divided by the new 125% total capacity of the combination of the two filters = 80%).
 The filter with the 25% capability will take the remaining 20% of the ammonia. If the smaller filter is anoxic which removes ammonia without producing nitrate and it is only doing that with 20% of the ammonia then 20% of the ammonia will not be converted to nitrate.
 In this scenario it can be very roughly estimated that the total output of nitrate will be reduced by 20%.

It’s possible to make some very rough estimates as to the total amount of the ammonia produced in the pond that will be drawn into the baskets and how much will go into the conventional biofilter.

If the total amount of ammonia going into a conventional biofilter can be reduced by, say, one-half then its nitrate output will also be reduced by almost a half.

The biocenosis baskets will remove the other half of the ammonia without leaving any nitrate. This combination of the two systems will mean that the total nitrate level in the pond will also be, very approximately, reduced by about a half.

The baskets also will remove some of the nitrate produced by the conventional filter which makes it difficult to take a mature conventional biofilter and estimate the overall effect of adding to it less than the full number of baskets that would be needed for full stand alone anoxic filtration.

.....if the number of baskets that would be required for complete stand-alone anoxic filtration should be, say, twenty but only five baskets, (25% of that number) are added to an existing conventional biofilter then the overall reduction in nitrate will be in the region of 20%.

For example; if the number of baskets that would be required for complete stand-alone anoxic filtration should be, say, twenty but only five baskets are added then the reduction in nitrate level can be expected to be in the region of 20%. It should be obvious that just adding a few baskets to an existing mature nitrogen cycle biofilter will not remove all the nitrate being produced by the existing filter.

I include that information because, as stated above, questions are often asked about whether one or two biocenosis baskets will reduce the nitrate level in a pond where the hobbyist is experiencing high levels. The answer is that full anoxic filtration, correctly implemented, will result in very low levels of ammonia, nitrite and nitrate; adding a few baskets to an existing conventional filtration system will obviously help reduce the nitrate level in the koi pond but there should be no unrealistic expectations as to their performance. It must be stressed that these figures are only guidelines, not exact values, but it will allow potential users to assess whether adding some biocenosis baskets to an existing filter system will produce a worthwhile reduction in nitrate.
Anoxic filtration - bug
How much oxygen does one basket remove?
Conventional filtration is called an aerobic process and is often described as a process where ammonia is oxidised, first into nitrite and then further oxidised into nitrate. As far as a chemist is concerned, that is a correct way to describe what happens to ammonia inside a conventional biofilter but koi keepers aren’t usually chemists and so the meaning of that description may not easily be apparent.

......The anoxic filtration pond doesn’t make any significant demand on oxygen and so the water coming out of an anoxic filter pond has the same dissolved oxygen concentration as the water that entered it.

The oxygen used is 4.3 times as much as the ammonia that gets converted!

The bugs in conventional biofilters have no option in this; if there is either insufficient oxygen or insufficient carbonate or dissolved carbon dioxide available, the process literally stops dead. They cannot begin to convert ammonia to nitrite and then to nitrate without oxygen so the ammonia just builds up in the pond water.

This is why we aerate conventional filters. If a conventional biofilter isn’t sufficiently aerated to provide enough oxygen for the activities of the total bug population, the bugs will take their oxygen from the dissolved oxygen in the water flowing through the media.

Anoxic filtration - conventional aeration 200
Conventional filtration requires so much oxygen that biological filter bays should be aerated to prevent the bugs depleting the dissolved oxygen level returning to the pond

That means that the water leaving an insufficiently aerated conventional filter on its return journey to the pond will be depleted in oxygen. The worst case example would be with a very active biofilter due to a high feeding rate combined with a relatively slow turnover rate. The water returning to the pond might, not only have a significant level of ammonia and/or nitrite but could have a dissolved oxygen level that is not sufficient for koi.

Unlike the bacteria that live in conventional filters, the bugs that live in the anoxic environment of a biocenosis basket are facultative anaerobes.

As described earlier, they still require oxygen but, in the anoxic environment in the centre of the basket (i.e. low oxygen – only 0.5 mg/L to 2 mg/L), the overall effect of all the bug activity is to take their oxygen from molecules of nitrate.

Anoxic filtration converts ammonia (NH4) directly to nitrogen gas (more correctly called dinitrogen N2) whilst simultaneously reducing any residual level of nitrate in the pond.

An anoxic filtration pond doesn’t make any significant demand on oxygen and so the water coming out of the anoxic filter pond has the same dissolved oxygen concentration as the water that entered it.

If biocenosis baskets only remove NH4, where does all the NH3 go?
When ammonia is excreted from the gills of a fish it’s initially in the toxic NH3 form. Some of it then becomes NH4 which is almost non-toxic. The amount of NH3 that changes to NH4 is determined mainly by the pH and, to a lesser extent, the temperature. What may be more difficult to understand is that the amount of each is a percentage of the total amount and is therefore dynamic, not fixed.

Biocenosis basketIf this seems to be a complicated relationship to understand you might try thinking about it this way:

Assume, for easy numbers, there are only a total of 100 ammonia molecules in a pond. If the pH and temperature is such that they split into 50% of each, there will be 50 NH4 and 50 NH3.  If you now take away two of the NH4 molecules, that would leave 48 NH4 and 50 NH3.

But, in this example, the 50 – 50 balance of the two forms dictates that there has to be 50% of each, so one of the NH3 molecules will spontaneously pick up a hydrogen ion (H) and become an NH4.

Science panel

Other percentages
In the example used in the main text a 50 – 50 split was used in order that the numbers could be easy to follow. In that example, each time two NH4 molecules disappeared into a basket an NH3 molecule converted to NH4.
 In other words, with a 50 – 50 split, every time the NH4 count was reduced by two, the NH3 count would reduce by one.
 If the pH and temperature caused a different percentage split between NH4 and NH3 the principle would remain the same but the numbers will be different.
E.g. 80% NH4 and 20% NH3 (or 4:1 ratio)
 If the pH and temperature dictated that there would be far less NH3 than NH4 such as 80% NH4 and 20% NH3, then one NH3 would spontaneously convert to NH4 every time five NH4 molecules disappeared into a basket.
E.g. 20% NH4 and 80% NH3 (or 1:4 ratio)
 If there was far more toxic NH3 than NH4 the situation would change.
 Every time five NH4 molecules disappeared into a basket a total of four NH3 molecules would spontaneously convert to NH4.
You can prove this
 Try these two examples with one hundred matches split into two piles with one pile representing NH4 and the other pile representing NH3. Keep removing the appropriate number of matches from the NH4 pile then convert the corresponding number of matches from NH3 to NH4 by shifting them from the NH3 pile to the NH4 pile.
 As the numbers of matches in both piles reduce, the percentages or ratios between them will always be the same.

If you keep taking away two NH4 molecules at a time, each time you do, the 50% relationship will only be satisfied by one of the NH3 molecules picking up an H and becoming an NH4.

If you follow that example to its conclusion, each time you reduce the NH4 number by two you will reduce the NH3 number by one until both the NH3 and the NH4 form of ammonia have been removed from the pond water.

So the fact that biocenosis baskets only attract the relatively non-toxic NH4 molecules into them doesn’t mean that they have no effect on the toxic NH3 molecules. In that example, in order to keep the numbers simple and easy to understand, I've used only 100 molecules and a 50% split.

However, you can work the same example with any percentage and also scale up the number of molecules to the actual astronomical numbers of individual ammonia molecules that you would find in a koi pond. The principle is the same; removing NH4 molecules also removes NH3 molecules because the percentages of each are dictated by pH and temperature.

Any imbalance in that relationship can only be satisfied by toxic NH3 molecules spontaneously converting to the non-toxic NH4 form which are then removed by the baskets.

The layout of the original pump fed system
Fortunately, building an anoxic pond is far easier than understanding how it works. In Kevin Novak’s original pump fed design, (figure 4), water is pumped from the main pond into the anoxic pond. In order to prevent the flow of water from disturbing the baskets, it enters through a simple diffuser. Figure 4 also shows Kevin’s original design for a diffuser which essentially consisted of a perforated pipe in a basket full of bio balls but any other design could be used if preferred.

The water then returns back to the main pond by gravity. A suitable depth for an anoxic pond is about 24 inches (600 mm) deep but it can be any convenient depth or shape that is large enough to allow approximately one basket per adult fish. For clarity, only three baskets are shown in the diagram.

Anoxic filtration - original pump fed design

Anoxic filtration - gravity fed design

Anoxic filtration - removable pipe 220Can anoxic ponds be gravity fed?
In the UK, hobbyists tend to have a preference for gravity fed filter systems as opposed to pump fed ones. It’s possible to modify the original pump fed design to a gravity fed system for those who don’t like pump fed systems or who want to modify an existing gravity fed system (see figure 5). As in the pump fed system, the water should be diffused as it enters the anoxic pond. One way to achieve this would be to extend the 4” bottom drain pipe above water level and to drill around 100 x ¼” (6 mm) holes in it.

If a plain socket is fitted as the 4 inch pipe enters the anoxic pond, and the perforated section of pipe isn’t glued into it, it will be removable and it can be replaced by a section of plain pipe with an end cap which can be used as a simple shut off valve for when the anoxic pond is being drained for maintenance.

Anoxic ponds fed by an airlift
An old principle that is being given a new lease of life as if it was a new invention is the airlift. Apart from being the oldest and, arguably, the most common method in use today for drawing water through under-gravel filters in aquaria, the airlift seems to have been pretty much ignored as a means of moving water in koi ponds, at least in the UK. However, rising energy prices are prompting more and more hobbyists to seek less expensive ways to move water from the koi pond to the filter system. The air lift is far more energy efficient than even the most economical pond pump and therefore is an ideal solution to reducing an energy bill for moving an equivalent volume of water per hour.

Apart from the altruistic benefits of a reduced carbon footprint for the good of the planet that we will be leaving our grandchildren, there is the more immediate benefit of reduced energy bills.

Anoxic filtration - airlift design

Figure 6 shows how an airlift could be used to draw water from the main pond and figure 7 shows the simplicity of the basic airlift itself. They are so adaptable that they can be made to suit individual situations so there isn’t a standard design.

It’s possible to design an airlift but that would involve solving several equations such as this one which is just for the minimum air flow. After solving that, there are more equations that will have to be solved for the other airlift parameters such as air stone depth, pipe diameter etc. Clearly designing from scratch isn’t practical.

Airlift equation_opt04

The alternative is to copy a design that is known to give good performance or to build one according to general principles and then experiment with the airflow to the air stone to find what rate gives the optimum water flow. It should be noted that the highest airflow does not always produce the greatest water flow since there is an optimum bubble to water ratio in the riser pipe for any particular length of riser.

In addition to experimenting with the rate at which air is fed to the air stone, its depth can also be varied to see which combination gives the best performance.

Anoxic filtration - airlift
Figure 7
Detail of airlift pipe

Science panel

Does an air lift lift?
 Air lifts can easily raise huge quantities of water slightly above the normal water level but, as the lift height above water increases, the flow rate decreases very rapidly. Figure 7 shows a simple air lift design suitable to supply water to an anoxic pond.
 The distance from the air stone to the water surface when the air is turned off is called the submerged length. It’s called the submerged length whether or not it is all inside the tank or pond and actually submerged or outside it because, when the air is turned off, the water levels will equalise and the inside of that part of the tube will be submerged regardless of whether any part of the outside of it is in water or in the dry.
 For an air lift to work, the submerged length must exceed the lift height.
 The greater the submerged length, compared to the lift height, the greater will be the flow.
 General guidelines for a DIY air lift are that the submerged length should be at least twice the length of the lift height and generally as long as is possible up to a maximum length of around two metres.

For a practical airlift, the diameter of the riser pipe can be as little as 2” but 4” pipe would provide a greater water flow and lift height for any given amount of air and so would be a better choice.

The riser pipe in the gravity fed version (Figure 5) was perforated in order to diffuse the water flowing into the anoxic pond so that the flow doesn’t disturb the baskets. The riser pipe for an airlift should not be perforated.

Since the airlift discharges at the surface of the anoxic pond there shouldn’t be too much disturbance to the baskets in the anoxic pond. However, if a powerful airlift should create too much surface disturbance, a tee can be fitted onto the riser pipe. This will help to spread the flow in two opposite directions and, if it hasn’t been glued, it can be turned in order to direct the flows so as to find the position that will reduce any residual disturbance as much as is possible.

Airlifts are not very well named in that, unless they have been specifically designed, they are very poor at lifting water. The arrangement shown in figure 6 will be a suitable design because it feeds back to the main pond via a 4” pipe. The anoxic filtration design shown in figure 4, where the water is lifted into a raised anoxic pond and flows back via a waterfall, will only be suitable for the pump to be replaced by an airlift if the height of waterfall doesn’t exceed about six inches.

The diagrams shown above are simplified to show only the principles of each installation; they don’t show such details as maintenance valves. If the airlift riser pipe terminates just above the normal water level in the anoxic pond, water cannot flow in from, or back to, the main pond when the air lift is stopped for maintenance. This won’t have a significant effect on the airlift performance but removes the need for a valve to isolate the inlet pipe connecting the main pond via the airlift to the anoxic pond. However, the outlet pipe back to the main pond will need an isolation valve.

Combined anoxic filtration and fry growing on pond?
Why not? There is no reason why small fish cannot be kept in the anoxic pond other than the limited amount of room and the fact that koi, if they are large enough, will want to try to excavate the baskets to see if they can find a morsel of food. With just those two constraints, the pond could be used as a permanent home for small fish or even for hatching eggs and growing the fry large enough to transfer into the main pond without the risk of them being eaten.

Those hobbyists who have an interest in spawning in their pond and use spawning ropes can simply transfer the egg coated ropes to the anoxic pond and allow them to hatch.

Those hobbyists who have an interest in spawning in their pond and use spawning ropes can simply transfer the egg coated ropes to the anoxic pond and allow them to hatch.

In the UK there is a developing interest in buying recently hatched fry from reputable breeders who have spawned high grade koi but are naturally unable to raise all the fry that have successfully hatched. These breeders will obviously have already picked out the fry that they judge will develop best but are happy to sell bags of the other fry even though there is a good likelihood that many of them will develop into excellent specimens since these would otherwise have to be destroyed anyway.

Quarantine procedures are always recommended when adding new fish to a pond but, after the quarantine period, the fry could be transferred into the anoxic pond and allowed to develop. Some thought will have to be given to ensuring that very small fry aren’t sucked into a pump or swept over a waterfall but a fine mesh screen or sheet of Japanese Matting could be strategically placed to avoid that risk.

Floating baskets and unusual locations
Biocenosis baskets don’t have to be in a separate anoxic pond if space is limited. They will work just as well in the main pond in floating baskets or in baskets suspended at the surface if this is done in such a way that inquisitive fish cannot excavate them. They can also be put into water features such as streams or rock pools but there are restrictions as far as water flow past them is concerned. They need a non-turbulent water flow so as not to wash the cat litter out of the baskets, either from the top or through the holes in the sides. Baskets in ponds can be protected from being excavated by putting them in the legs of stockings or tights (pantyhose) and tying the tops loosely around any plant stems.

Anoxic filtration is simple, inexpensive and very efficient if you follow the rules but doomed to failure if you start altering the design parameters without having understood the principles.

Water doesn’t flow through a biocenosis basket as it does through a media bed in conventional filtration. There is a possible problem if one side of a basket faces a strong flow of water and the opposite side is downstream of the flow. The pressure of water onto the side facing the flow might cause freshly aerated water to penetrate into the basket. Worse still, the differential pressure across the basket might even force water to flow through it. The first scenario would make the basket less efficient, the second would prevent the anoxic environment forming in the centre of the basket. Anoxic filtration is simple, inexpensive and very efficient if you follow the rules but doomed to failure if you start altering the design parameters without having understood the principles.

Anoxic build 24 (380 px)

If more baskets are required they can be stacked either randomly, as in this example, or even as a complete new layer. The only constraint is that each extra basket should be spaced from the one below so that it doesn’t rest directly on it. This is to allow water to circulate freely around all surfaces of all the baskets.

Are there any drawbacks?
There are no drawbacks but one point is worth careful consideration. Settlement will occur in the anoxic pond and it will eventually need to be emptied or flushed to waste just as any other settlement chamber. In order to keep the drawings as simple as possible, I have left out details of prefiltration and a drain to make emptying easier.

A sieve is a suitable pre filter for the gravity system. A simple way to close off the main pond when a gravity fed anoxic pond is being emptied would be to make the perforated section of pipe removable and have a suitable length of un-perforated pipe with an end cap that can replace it whilst emptying.

If an airlift is used to move water from the main pond and the riser pipe terminates just above the water level in the anoxic pond it, won’t be necessary to have a valve, or the removable riser pipe as described above, to isolate the anoxic pond from the main pond for maintenance purposes since water won’t flow from one pond to another when the air flow is stopped.

And the advantages?
Apart from reducing ammonia and nitrate levels, the system also reduces phosphate levels too. A molecule of phosphate (PO4) consists of one atom of phosphorous (P) joined to four atoms of oxygen (O). Phosphorous is essential for plants (including algae) since it’s required to build the plant’s DNA and is also needed to make the energy molecule, ATP (adenosine triphosphate), which is necessary in order for the plant to convert carbon dioxide, water and oxygen into the carbohydrates that it will need to build new cells and grow. By drawing in and destroying phosphate in a biocenosis basket it is removed from the pond water where it could encourage algae but the phosphorous is still available to provide nutrients for any plant in it.

Molecule - phosphate
A molecule of phosphate (PO4) consists of one atom of phosphorous (P) joined to four atoms of oxygen (O)

Phosphate in the koi pond

All living creatures and plants need phosphorous, it has several functions. All cells need it to make and to copy their RNA or DNA in order for them to divide. It’s a necessary component of an essential molecule called ATP (adenosine triphosphate). A single molecule of ATP contains three groups of phosphate molecules and these are what allow ATP to give all living cells the energy to go about doing whatever that cell is supposed to do.

The multiple ways in which ATP is used by cells is really a topic for a separate in-depth article but these are a few of the ways in which ATP is necessary for life:

ATP gives muscles the ability to contract, it gives the digestive system the power to digest, it powers the immune system and allows scar tissue to form and heal wounds.

Science panel

Simplified diagram of photosynthesis
Plant  photosynthesis

 When light shines onto the chlorophyll in special cells called thylakoids that are in plants or algae, it causes water molecules (H2O) to be split apart.  The oxygen is a waste product from that reaction and is released which explains how plants produce oxygen during photosynthesis.
 They don’t produce oxygen by design. Oxygen is simply their industrial waste which they carelessly dump into the environment!
 Plants only need the hydrogen from this light reaction and it is then combined with the phosphorous in phosphate and used to create ATP and another molecule called NADPH.
 These are very high energy molecules that power a separate process which converts carbon dioxide into the sugars and carbohydrates that plants, algae and blanket weed/cyanobacteria need to build new cells and grow.

Phosphate is also necessary to allow plants to photosynthesise and so reducing its level in the koi pond severely limits the ability of algae or cyanobacteria / blanket weed to grow and to multiply.

How phosphate gets into a koi pond
Since all living cells need phosphate, it will be found in all koi foods that are made from natural basic food ingredients. Phosphate will also be in the fresh food treats that we feed to our koi. Prawns, for instance, are particularly high in phosphates, containing as much as 230 mg per 100 grams.

Fish extract from their diet the amount of phosphate that they need and their kidneys excrete any excess. Phosphate excreted in this way, along with any phosphate that leaches from foods before it is eaten, slowly builds in the pond water.

Phosphate is difficult to remove from a pond. Where ponds only have conventional biological filtration, it will only be removed during water changes. Just as in the case of nitrate removal, a 10% water change will reduce the levels of both the nitrate and phosphate by 10% but the levels will begin increasing again with the next feed.

How anoxic filtration removes phosphate
As described earlier, when facultative anaerobes are in an anoxic environment, they switch their mode of respiration and can take the oxygen they need from nitrate.

Since phosphate (PO4) consists of one atom of phosphorous (P) joined to four atoms of oxygen (O) the bugs are also able to take their oxygen from any phosphate in the water.

Additionally, baskets that contain plants will take phosphate directly through their roots.

Building an anoxic filtration system

The anoxic system is inexpensive to build and is adaptable. It will fit anywhere because it can be built to fit whatever space is available. The only constraint is that there should be about one basket per adult fish but, even in this, there is flexibility.  If ever it is found that more baskets are needed for an increased stocking level, and space is limited, an extra layer can be stacked on top of the bottom layer as long as they are spaced so that the bottom of one basket doesn’t rest directly on top of the one below. This will ensure that water can still flow past all surfaces of all baskets.

Anoxic filtration - spaces around baskets clrday

Small spacers allows water to flow past all surfaces of the baskets

A basket where water is able to flow past all six surfaces, (the four sides plus the top and bottom), will be able to remove ammonia more efficiently than a basket where water can’t flow past one or more of the surfaces. The gap doesn’t have to be great and it can be seen from the picture below of an anoxic pond that, in order to minimise the footprint of the pond, the baskets have been placed with their tops actually touching.

However, since typical baskets taper toward their bottoms, even if the tops are touching, there will still be adequate spaces between them for water to freely circulate around all sides. If small spacers are placed underneath the baskets, they will allow water to circulate underneath them too.

Stacked baskets can reduce the anoxic pond footprint
Anoxic ponds can be of any shape or size to suit available space and the depth of 60 cm (24 inches) is only a suggestion that will suit a wide range of common aquatic plants. The depth isn’t important to how efficiently the baskets will function; they will work at any depth that is convenient to the hobbyist from being just covered by water to a depth of one metre or more.

Deep anoxic ponds will allow the baskets in it to be stacked in order to get the maximum number of baskets into any given surface footprint. However, if a deep anoxic pond is considered, thought should also be given to the ease of future maintenance.

As a guide, there should be about one basket per adult fish for the system to provide stand-alone anoxic filtration, just putting one or two biocenosis baskets somewhere in the pond may not have any significant effect.

As a guide, there should be about one basket per adult fish for the system to provide stand-alone anoxic filtration, just putting one or two biocenosis baskets somewhere in the pond may not have any significant effect. The optimum size of the biocenosis basket is 30 cm x 30 cm x 20 cm. It’s possible to use smaller baskets but they will have less room inside them for anoxic conditions to develop so they won’t be as effective as the optimum size.

Larger baskets would allow for a larger anoxic zone to develop but will become very heavy once they have been submerged. A point that should be remembered is that each basket may have to be lifted out to allow for occasional maintenance of the basket itself or cleaning any leaves that may blow into the anoxic pond or sediment that may accumulate in it. A leaf net over the anoxic pond will reduce maintenance.

Anoxic pond without plants
A typical anoxic pond without plants

This example of a typical anoxic pond contains 22 baskets that are 30 cm x 30 cm square by 20 cm deep and which are contained in a pond that is 1.8 m long by 1.2 m wide. The water depth over each basket is around 40 cm.

These dimensions aren’t crucial, they are only for guidance. This particular pond is just an example of a compact way to house 22 baskets.

Any other convenient shape would be just as suitable as long as all the baskets are completely submerged and that water can flow gently through the pond without disturbing the contents of the baskets.

Avoiding settlement in the anoxic pond
Whichever filtration method is used, conventional or anoxic filtration, a turnover rate of once per hour or greater is recommended. The faster ammonia is removed from where the koi are excreting it and presented to the biofilter the better because the overall effect will be to reduce background levels of ammonia in the main pond.

Since nitrifying bugs rapidly colonise any wet surface where there is also a source of ammonia, every surface in a koi pond will provide a home for the ammonia bug (nitrosomonas). Their waste product, nitrite, will then provide a nutrient source for the nitrite bug (nitrobacter) which will also colonise every wet surface in the pond. The result is that there will always be a degree of the nitrogen cycle conversion of ammonia to nitrate in koi ponds.

Anoxic filtration - sieve 1With conventional filtration, this isn’t a problem because the biofilter will also be converting ammonia to nitrate so any similar action taking place in the main pond will assist the biofilter not oppose it.

Where anoxic filtration is employed, the ammonia is removed from the system without leaving nitrate as the end product. Therefore a quicker rate of removal of ammonia from the main pond and the consequential reduction of background ammonia in it is not only of benefit to the koi but it also reduces, as far as is possible, the nitrogen cycle taking place in the main pond. This allows the majority of ammonia to be removed and dealt with in the anoxic without nitrate being produced as an end product.

Accordingly, the recommended flow rate through an anoxic pond is as fast as is possible without disturbing the contents of the biocenosis baskets. Even with a relatively high pond turnover rate, the rate at which water flows through the anoxic pond will be far slower than the rate at which it passed through the pipework leading to it. Just as high winds will be able to blow over dustbins but a gentle breeze will only be able to blow paper litter around, water travelling at higher speeds is able to carry more silt, and heavier silt, than water travelling at slower speeds. The slower rate of flow through the anoxic pond than in the pipework leading to it inevitably will lead to it being unable to carry so much suspended silt so there will be some settlement of the suspended particles that the water had been carrying. Provided any settled silt isn’t allowed to build to a depth of more than a few millimetres and it doesn’t contain any organic debris such as fallen leaves, it will become inert and will not lead to a subsequent deterioration in water parameters.
Rotary drum filter
One recommended addition to an anoxic system is a pre filter. Installing a good pre filter before the anoxic pond will remove most of the suspended particles. This in turn will substantially reduce the amount of silt that will subsequently accumulate in the anoxic pond although, even with a pre filter, there still will be some settlement. Any good pre filter, including brushes or a settlement chamber, can be used before the anoxic pond.

Trials by Kevin Novak and the users of anoxic filtration in the UK have shown that two types, a sieve or a rotary drum filter are particularly effective at removing suspended particles which will reduce the frequency at which the anoxic pond will have to be cleaned. Another biological filter that has been used by another UK user to good effect as a prefilter is a pump fed bead filter.

Plumbing is simple for either the original pump fed system, as is shown in figure 8, or the gravity system in figure 9. Other pre filter arrangements such as rotary drum filters, brush chambers or settlement chambers can be used especially if they already exist and the anoxic filter system is being used as a replacement for an existing conventional filter system. The plumbing shown in the diagrams will remain substantially the same if other types of pre filters are used instead.

The sieve and the associated plumbing in figure 8 can be replaced by a rotary drum filter or any suitable pre filter or settlement chamber that is compatible with being pump fed and which can then flow by gravity via the anoxic pond back to the main pond.

The sieve in figure 9 can be replaced by any pre filter that is compatible with being gravity fed and with an output that can be pumped to the anoxic pond.

Anoxic filtration - pump fed sieve

Anoxic filtration - gravity sieve

Removing the residual settlement that does collect in the anoxic pond

Earlier diagrams have deliberately been kept as simple as possible and have shown only the essential principles of how the anoxic filter system works and integrates with standard design koi ponds and equipment.

Interrupting the anoxic environment will only be a significant factor if the biocenosis baskets are left to drain for too long during cleaning or transferring to another pond …..
.…. I don't see a problem arising from cleaning the anoxic pond provided it isn't done every week.

The diagrams in this section show the extra detail of drains and associated waste pipework which can be periodically used to flush away any remaining settlement that does collect. A thin layer of fine silt with no accumulated fallen leaves or other organic matter will become inert so, with a pre filter, the anoxic pond won't need cleaning too often.

When cleaning the anoxic pond, a point to bear in mind but which shouldn’t be exaggerated is that, if the anoxic pond is completely drained or if a biocenosis basket is removed from water, the oxygen depleted water in zone C obviously will drain out. The water in the basket won't immediately drain completely; some oxygen depleted water will remain inside and will only drain slowly if the basket is left out of water for too long. When the pond is refilled or the basket is resubmerged, the anoxic environment will be removed as aerated water floods back in.

The anoxic environment will soon be restored but, depending on how much water had drained out and therefore how much aerated water goes back in, the interruption will either be a short blip or, in the case of a complete dry out, it may take much longer. Interrupting the anoxic environment will only be a significant factor if the baskets are left to drain for too long during cleaning or transferring to another pond so that all the oxygen depleted water drains out. I don't see a problem arising from cleaning the anoxic pond provided it isn't done every week.

As described earlier, in order to minimise the pond footprint, baskets may be positioned close together and possibly even with the tops of tapered baskets touching. However, if space permits, it might be prudent to have them slightly separated so that a hose may be used while the pond is draining to help wash the silt towards the drain in order to speed up the process.

Earlier diagrams have been deliberately kept as simple as possible and have shown only the essential principles of how an anoxic filter system works and integrates with standard design koi ponds. The diagrams in this section show the extra detail of drains and associated waste pipework.....

If a hose is used, care should be taken not to run it too fast or direct the water directly onto the sides of the baskets in order not to disturb their contents. To help achieve this, a length of pipe could be fitted to the hose to help direct water at the floor in between the baskets.

Any pipe will do for this purpose but 15 mm (½”) copper pipe is a suitable size and is conveniently light and strong. Despite sensible warnings about not using copper pipe in koi pond plumbing systems, a metre or so of copper pipe used in this way will not introduce copper into the water.

When cleaning the anoxic pond, it isn’t necessary to drain it completely, it’s only necessary to open the drain valve for as long as it takes to clear the floor but, if a great deal of water is lost in the process, the water won’t have been wasted, it will count toward that week’s quota of water changes.

The pump fed anoxic filter system
Figure 10 shows the waste valves and associated plumbing for a pump fed anoxic filter system with a prefilter. The floor of the anoxic pond is shown slightly sloped towards the drain because this will help remove any settled silt when the pond is being flushed.

Anoxic filtration - pump fed sieve + waste

In normal operation valve 1 is closed and is only opened to flush the anoxic pond. The waste pipework to the anoxic pond should be at least 2”. Good quality 4” valves are expensive but, if the budget allows, clearing the floor will be far more effective with the larger size since the increased flow to waste will create a greater “suction”.

Unless the prefilter requires a permanent connection to waste, valve 2 is normally closed and only opened to flush it. The waste plumbing to the prefilter should be sized as per manufacturers’ specifications for the type that is being used.

The gravity fed anoxic filter system
Figure 11 shows the waste valves and associated plumbing for a gravity fed system with a prefilter.

Anoxic filtration - gravity sieve + waste

As described above, the floor of the anoxic pond is slightly sloped towards the drain to assist flushing away silt. The waste plumbing and operation of the two valves is also the same as above with valve 1 normally closed except when flushing silt from the anoxic pond and valve 2 normally closed and only opened to flush the prefilter unless it requires a permanent connection to waste.

It’s advisable with conventional gravity fed systems that a valve should be included that will allow the bottom drain pipework to be regularly flushed to waste in order to flush out any silt that has settled in it. Valve 3 has been included for this purpose and, as is usual with these valves in a conventional system, it is kept closed during normal operation and opened, on a regular basis, as part of normal pond maintenance regardless of whether silt can be seen in the anoxic pond.

The airlift fed anoxic filter system
Figure 12 shows the waste valves and associated plumbing for an airlift fed system.

Anoxic filtration - airlift design + waste

As with the above options the floor of the anoxic pond is slightly sloped towards the drain to assist flushing away silt and valve 1 is normally closed except when flushing silt from the anoxic pond. In order to avoid the build up of settled silt in the bottom drain pipework to the airlift, valve 2 has been included and is opened, on a regular basis, as part of normal pond maintenance regardless of whether on not silt can be seen in the anoxic pond.

A novel approach to building an anoxic pond

This step-by-step build of an anoxic pond was documented and kindly supplied by Brian Woodcock.

Anoxic build 1
Paving slabs were cut with a 4" disc cutter and laid as a base.

Anoxic build 2
Construction of the wooden framework for the anoxic pond is 4 x 2 inch timber.

Anoxic build 3
One side and the two ends of the framework are quickly finished and in position.

Anoxic build 4
The framework is completed. Note the supports for the raised floor which allow room for the waste pipe.

Anoxic build 5
The sides and the floor of the anoxic pond are made of ½ inch marine ply.

Anoxic build 6
The total cost so far for the completed framework and ply lining is (GB) £180 [2014 prices].

Anoxic build 7
The diffuser that will spread the flow into the anoxic pond is completed.

Anoxic build 8
The diffuser is fitted.  When the liner is installed it will be turned through 45°.

Anoxic build 9
Under the floor of the pond, a ball valve was used on the waste line as they are more reliable than a cheaper slide valve.

Anoxic build 10
A boxweld liner (a heat welded prefabricated 0.75 mm EPDM or Butyl liner) provides the waterproof finish for the anoxic pond.

Anoxic build 11
With the liner in place the return inlet, (feed from the pond), was fitted through the ply sides and sealed to the liner.

Anoxic build 12
After fitting the return inlet, the fitting for the drain is fitted and sealed in the same way.

Anoxic build 13
The diffuser is turned through 45° to spread the flow across the whole of one end and provide a more diffuse flow.

Anoxic build 14
The anoxic pond copings match and complement the existing pond copings so as not spoil the look of the main pond.

Anoxic build 15
The water chute that will return water back to the main pond also matches the copings.

Anoxic build 16
Seven biocenosis baskets down nineteen to go.  Note that lengths of pipe are used to space baskets are from the bottom.

Anoxic build 17
Three hours later all the biocenosis baskets are filled.  The layer of stones prevents water flow disturbing the cat litter.

Anoxic build 18
A major goal was that it shouldn’t look like an add-on.  Plants will be added next to make it look less stark.

Anoxic build 19
Before the anoxic filter was started the nitrate level was over 80 mg/L.  Just ten days later it was already starting to fall.

Anoxic build 20
Before the anoxic filter was started the nitrate level was over 80 mg/L.  Just ten days later it was already starting to fall.

Anoxic build 21
After two months nitrate levelled at 30 mg/L despite feeding 300 g per day to 21 koi and a bead filter producing nitrate.

Anoxic build 22
A lily that was very small when bought is already producing its first bud.

Anoxic build 23
A few hours later the first water lily opens and compliments the luxuriant plant growth.

Anoxic build 2402
To clean, after emptying, a hose is used to rinse mulm towards the drain. Total time from start to finish - 20 minutes.

Anoxic filtration in the UK

The anoxic filtration system was design and developed by Dr Kevin Novak PhD and in the US, where Kevin lives, all cat litters are pretty much the same. In the UK there are many different kinds of cat litter type products and manufacturers sometimes change brands and ingredients so those intending to use anoxic filtration will have to look around to see what’s available in their area.

Cat litter 01

Cat litter 02

Correct types of cat litter

These are some commonly available brands that have retained their structure in tests and which have been used successfully but you should check that the composition hasn’t changed before buying large quantities. Obviously don’t try to use any cat litter that is made from recycled paper or wood pulp or which promises to naturally decompose in an environmentally friendly way when disposed of after the cat has used it.

The panel on the left shows the type of granular clay that you should be looking for. The colour isn’t important and it need not be little round granules but it mustn’t be so fine that it packs down like sand. Biocenosis baskets need to have open pathways in between the granules in order to allow the ammonia to be drawn into the centre. The panel on the right shows types of cat litter that should not be used.

The cat litter should be thoroughly rinsed in order to remove any dust but litter that contains excessive amounts of dust won’t be suitable since there may be very little of it left when all the dust has been rinsed away.

Clumping cat litter03
Clumping cat litter

Recycled paper pellets02
Recycled paper pellets

Odour control litter02
Odour control pearls

It’s preferable not to use scented cat litter but if you can’t find a suitable unscented one then a scented litter can be used as long as it doesn’t contain either scented or odour control pearls and provided the perfume is thoroughly rinsed away. Products described as “Clumping Cat Litter” are invariably useless because, when they get wet, the granules clump together. This is perfect when cat litter is used for its original purpose but if it is permanently submerged it will just become a soggy mush.

Some people have found that a product called “Oildri ©” or plain “Fuller's Earth ©” are consistently suitable types of media.

Unless you have a cat or know someone who will appreciate an opened bag of litter that turned out to be ideal for a cat to mess in but no good for you, instead of buying a bag that looks promising, try to obtain a few grains from it, take them home and put them in a glass of water overnight. Squeeze them in the morning and if they go to mush, they're no good. If they seem ok, leave them for several days and try again.

When you find a product that retains its granular structure after being submerged for several days, make up a test basket without the laterite and submerge it for at least a week and then take it out and delve into the cat litter to make sure it still retains its granular structure. If it does then you have found a suitable product so you can add the laterite to that one and obtain sufficient cat litter of that brand to make all your baskets, complete with laterite, and use them straight away in full confidence.



Brands of cat litter on sale in the UK that have been reported as being suitable

Before buying large quantities of cat litter, even the brands shown below which have been tested and found to be successful, it’s advisable to first check its suitability in order to ensure that the manufacturers haven’t changed the type or the grade of clay used. This is easy to do using a small amount and the method described above.


Steetley Sophisticat Original
Steetley Sophisticat Original

Tesco odour control06
Tesco odour control cat litter

Tesco cat litter
Tesco low dust cat litter

Tesco premium cat litter
Tesco Premium cat litter


Laterite 135 230

Search around for the best price. Laterite can be found on the Internet including postage for the 1.6 kg size for less than £20 / US $33 / €25 [2014 prices and currency conversion].

This size will be sufficient for between four and eight baskets using about 200 gm to 400 gm per basket depending on the basket size.

As a guide, for effective stand-alone anoxic filtration you will need about one basket per adult fish.



Plants 204

There are no specific plants to use. As long as they will grow during the summer season rather than be a static constant size and will not develop a root system that will grow so large that they will burst the basket you can choose plants you like the look of and they will be fine.

Plants 403

Plants 103

Plants 504

Plants 303


An alternative to cat litter
There are places where a suitable type of cat litter simply isn’t available. Lio Fornellino (Dr. Franco), from Italy, was keen to test the performance of the anoxic filtration system. He searched throughout Italy for a suitable product but was unable to find one.

Lio teeshirtHe asked for help in locating an alternative on an anoxic filtration forum. One UK user of the forum bought a large consignment, arranged to take it to Dover (UK) and Lio drove from Italy to collect it. He tested the system under different scenarios and was so impressed with the results that, before his untimely passing in October 2012, he went on to promote anoxic system throughout Italy via his blog.

Since then, another possible source of the correct type of clay has been brought to my attention. It is used as an addition to bonsai soil and is slightly more expensive than cat litter although current UK prices for a 25 litre bag are competitive when compared to 2½ times the cost of 10 litres of a suitable cat litter. In the UK it is marketed as Biosorb, Ultrazorb or Danish Pink and anyone having trouble locating a suitable product in their area might consider looking at what is available via bonsai equipment selling sites on the Internet.

Anoxic filtration has been developed over a period in excess of twenty five years by Dr. Kevin Novak PhD. Whilst researching for an article about anoxic filtration in 2008 I found that Kevin had written a CD book about it so I emailed asking where I could buy a copy. Kevin insisted on sending me a copy FOC and even refused to accept the refund of the Trans-Atlantic postage from the US to the UK.

I wish to express my deep appreciation to Kevin for the copy of his CD and for the help he has given to myself and to others via his forum posts and his blog website. Thank you Kevin.

Kevin’s updated book is now available as a free download on iTunes:

Kevin also has FAQs and gives other useful information on his blog is on this link: