Koi Filtration

Koi, Japanese Koi fish, Butterfly Koi, Koi varieties, Koi identification pictures and everything on "HOW TO" information!

Koi Sale Prices
Other Websites
Koi Book Store
We Buy Koi
Koi Varieties
Koi Growth
Koi Weight
Koi Age
Koi History
Koi Taxonomy
Koi Sexing
Koi Breeding
Koi Can Hear
Koi Can Taste
Koi Can Smell
Koi Can Blush
Koi Can Overeat
Koi Nitrification
Koi Filtration
Pump Savings
Pump head
Koi Need Oxygen
Koi Shows
Koi Quality
Koi Feeding
Why Koi?
Koi Imports
Koi Pond Liners
Koi Pond Tours
Koi Pond Plants
Koi Predators
Koi Patents
Koi Equipment
Koi Services
Koi Store Pictures
Koi Water Quality
Koi Water Testing
Koi Health
Koi Pond Heating
Keywork Search
Koi Index

State-of-the-Art in Pond Filtration

By David A. Dec ©2002

Pond filtration requires two main components; both mechanical and biofiltration. First, the mechanical filtration needs to remove both the suspended particles and the debris that has settled to the bottom so the water is clear, and we can see our fish and plants. If not removed, this organic mulm can harbor dangerous organisms and chemicals that threaten the health of our pond.

Second, the biofiltration needs to change the toxic chemicals like ammonia and nitrites, to harmless or helpful chemicals like nitrates, oxygen, and nitrogen. This is done with biologically active bacteria species through the Nitrification process. 

Effective Pond Volume

A recent buzzword in pond circles is “Effective Pond Volume”. People have found that selecting a filter based on how many gallons their pond holds is not at all accurate. Since biofiltration is dependent on the ammonia load from the pond’s fish, biofilters needs to be sized according to the amount of fish wastes produced each day in the pond. This depends on the quantity of foods they consume each day.

Selecting a swimming pool filter is done by matching flow rates. However, the proper way to select a filter for a pond is to determine the maximum number of fish you are planning for the pond, and the weight of food you will feed them daily, plus the amount of algae and other natural pond food they eat. Typically the amount of food Koi eat is 1% to 3% of their weight. 

How Much Food

Then you select the filter that will process that weight of food per day. However, this is not an easy task because many filter manufacturers’ feel compelled to match their competitors' exaggerated claims of their filters' capacities, which means the ability to process food to ammonia and nitrite levels of 0.0 to 0.1 parts per million (ppm). Unfortunately these manufacturers’ grossly overstate the amount of fish wastes that their filters will handle. We have done our own research into this and found most of their claims to be outlandish. Surprisingly if they were more accurate they would probably sell larger or multiple filter units to their customers, the result would be healthier fish, and everyone would be happier.

Don't forget the square area of the pond itself supplements the square area of the filter media, and can easily add 25% to 200% to the bacteria's total processing surface area.

Overfeeding the total square area for nitrification causes the ammonia and nitrite levels to increase to dangerously toxic levels, without frequent water changes. That is why many ponds are running ammonia and nitrite levels of 0.5 to 1 or more ppm, and their owners are wondering why their fish are sick and dieing.

Filters can also be combined in parallel or series depending on the application or the amount of food that needs to be processed.

Surface Area Efficiency

Our own research, as well as others’, has shown that filters process about 0.1 gram of food per day per square foot of surface area in the filter media, at maximum nitrification efficiency. In other words, they convert the toxic ammonia to nitrites, and then the toxic nitrites to nitrates via the nitrification cycle. Since the bacteria live on the surface of the pond walls and filtration media, a larger surface area in the filter means there is more room for a larger bacteria colony to do the biofiltration.

The surface area in a filter is the filter media’s surface area per cubic foot, times the number of cubic feet of media. A larger total surface area allows larger colonies of nitrifying bacteria to adhere to it.

The surface area in a filter is what makes it work. The larger the surface area the greater is its biological activity.

The most important consideration in purchasing a filter system is its surface area efficiency in dollars per square feet.


The surface area per cubic foot depends on the media. For instance, very fine sand with a diameter of 0.125 mm or .005 inches has 8,000 square feet of surface area per cubic foot. So a sand filter with 2 cubic feet of very fine sand has 16,000 square feet of surface area so it can process 1,600 grams or 3.5 pounds of food per day. Sand of this diameter in a filter results in a total cost about a nickel per square foot of surface area.

Sand with a diameter of 0.5 mm or .02 inches has 2,000 square feet of surface area per cubic foot; and results in a cost of about $0.20 per square foot of surface area.

Plastic Bead Filters

Plastic beads with a diameter of 5 mm or 0.2 inch have a surface area of 200 square feet per cubic feet or about 40 times less than sand. So a bead filter with the same 2 cubic feet of beads has only 400 square feet of surface area, so it can only process 40 grams or 1.5 ounces of food per day; much less than that for the sand filter.

For this reason when beads are used in filters the typical cost is $1.00 to $2.00 per square foot of surface area or 5x more than sand.

It is obvious that the media needs to have a very large surface area to host the nitrification bacteria. All the media up to this point have been solid, i.e. sand or plastic beads.

Hollow Media

A newer development is hollow media. If you took a miniature plastic “hollow drinking straw”, and formed internal walls inside it, you would not only have the external surface area, but also an internal surface area. They are 5 mm in diameter and vary in length from ¼” to ½”.

This media has 750 square feet of surface area per cubic foot. So a filter with 2 cubic feet of media will have 1,500 square feet of surface area, and will be able to process 150 grams or 1/3 pound of food per day. This is over 4 times more efficient than bead filters, but 1/10th of the sand filter’s surface area.

Its cost when used in a filter is about $0.75 per square foot of surface area or more than 3x sand.

The first disadvantage of this new media is the difficulty of manufacturing it. The units are obviously very small and have a very complicated design, which makes them very expensive at about $200/ cu ft.

The second disadvantage is they are made of styrene. You can test this by putting some media in water. If it sinks it is most likely styrene. Another test is to burn it since styrene burns with a black sooty smoke. Styrene oxide as a contaminant of styrene is known to be very toxic to bacteria, which would be a disaster to bacterial colonies for nitrification.

Filter Operation

It is true that sand filters have been used in almost every swimming pool in the USA. They have been proven and constantly improved after millions of installations. They are obviously very effective, inexpensive, and are very easy to clean.

In fact, they are almost self-cleaning. You just turn a valve to backwash them. This saves lots of time, inconvenience, labor, mess, and wear and tear on the pond fish, and it automatically provides the small but frequent water changes needed to remove the dissolved chemicals.

During normal operation the water flows in through the top valve, travels down through the sand or filter media where the debris is trapped, and flows into 6 to 8 perforated plastic laterals connected to a hollow stem, and then out through the top valve back to the pool. During backwash the water is re-directed down the stem and through the plastic laterals, flowing up through the sand, carrying the debris out through the top valve to waste.

One way to select a swimming pool filter is by flow rates. For instance, if you are planning a flow rate of 110 GPM or 6,600 GPH you will want one 36" diameter filter, or possibly split the flow between two 24" filters, which would be cheaper.

Filtration flow rates in gallons per hour vary with the size of the filters as follows:

Filter Diameter

Typical volume in cubic feet

Flow Rates in GPM

Flow Rates in GPH



8 - 11

480 - 660



12 - 18

720 - 1,080



19 - 28

1,140 - 1,680



29 - 34

1,740 - 2,040



35 - 45

2,100 - 2,700



46 - 68

2,760 - 4,080



69 - 100

4,140 - 6,000



101 - 165

6,060 - 9,900



166 - 269

9,960 - 16,140



270 - 360

16,200 - 21,600

Notice that two 36" filters have about the same performance as one 48" filter, and are a bit cheaper, but will have much less back pressure.

Sand filters are typically charged with 1/2 of their cubic foot volume, with 100 pounds of sand being equivalent to 1 cubic foot.


Another benefit of the "pressurized" sand filters is the ability to use the pump's suction line to operate a vacuum to clean the bottom of a pond. The vacuum hose typically plugs into a skimmer's suction line to the pump.

The valve is turned to “Waste” so the vacuumed waste does not go through the filter, but goes directly to the waste line.

This same vacuum hose can also operate a mechanical robot vacuum that automatically vacuums the pool's bottom. These robots are made for either concrete or EPDM liners.


With the pressure type filter the skimmer’s strainer-basket becomes the pre-filter. We often have to clean the skimmer strainer-basket of algae, leaves, and debris twice a day, and the pump strainer-basket at least twice a week. However, it is quite easy, you just remove the basket, hose it off, and replace it. It only takes a minute or two. You can use chlorinated garden-hose water since the basket is not part of the biofilter.

Problems with Sand

There are 2 main objections to sand filters for ponds. First and most important, they can plug up if:

  1. Loaded at 100% of the manufacturers’ recommendation for sand, which is about 1/2 full.
  2. Not backwashed at least once per week.
  3. Not backwashed with a powerful enough pump.

Second, some opponents say the water travels through it too fast to allow for effective biofiltration. They say the residence time is too short. However, they ignore the fact that the water makes many more trips through the media for a given time period, so the actual contact time per hour is about the same.

Large City Aquariums use Sand Filter

Most if not all large city Aquariums use sand filters. They know how to properly use them, and have found the efficiencies to be unsurpassed.

Pea Gravel

Not knowing how to properly use a sand filter some people tried replacing the sand with pea-gravel, which has a surface area of only 100 square feet per cubic foot, or 80 times less than sand. Needless to say pea-gravel is not the answer.  


Diameter mm

Area in square feet per cubic foot (ft2 / ft3)

Pea Gravel



Very coarse












Very fine



Very very fine



Bead Filters

Somebody noticed that the plastic-bead feedstock used by plastic injection molders had diameters much smaller than pea-gravel, but larger than sand. These solid plastic beads, at 3 to 5 mm, 1/8” to 1/5", have a surface area per cubic foot of 200 to 300 sq ft / cu ft, which is better than pea-gravel, but 30 to 40 times less efficient than sand. In other words, you might need 30 to 40 bead filters to match the biological efficiency of a single sand filter.

Plastic beads diameter in mm

area in square feet per cubic foot (ft2 / ft3)















In order to get more surface area in the bead filters manufacturers simply try to put more beads into the filter; since the beads have much less surface area per cubic foot. They jam it almost full; in some cases they fill 90% of the filters' volume with beads. This leaves little room for backwash turbulence to develop.

So even with the dramatically reduced efficiency the literature shows many bead filters are still plugging up: partly because of the overfilling, and partly for the same reasons listed above for sand filters; not backwashing often enough, and not backwashing with a powerful enough pump. 

So some manufacturers have added air blowers to try to reduce this plugging tendency. Unfortunately, when the beads clump up forming channels the air simply goes through the path of least resistance, the channels, which means it has no effect.

The manufacturing of “bead” filters was very simple and became quite popular. It consisted of buying standard sand filters at wholesale, dumping the sand, and inserting the polypropylene or polyethylene plastic beads, and jacking up the price 3-4X for sale to the pond industry. The drainage plugs were now referred to as “sludge removal ports”.

Then some very clever bead filter manufacturers, realizing they needed to add more value, added small UV lights, and compressed air-lines.

Bead Media Washout

During the backwash operation sand being heavier than water falls to the bottom of the tank, instead of flowing out through the valve to waste. However, the plastic beads being lighter than water float to the top, and since they are smaller than the valve-strainer's holes, they are washed out through the valve into the waste stream; so more and more beads are lost during each backwash operation. 

This limits the size of the beads being used; the smaller the beads, the greater the surface area for bacteria, but the more bead loss during backwash operations. The larger the beads the smaller the surface area for bacteria, but the backwash bead loss is reduced. So the bead filters are limited in efficiency. 

Pond Sand Filter Research

Our research focused on under-loading the sand filters, and backwashing them more frequently with higher pressures and flow rates, in order to take advantage of the greater food processing surface areas, while eliminating the chance of plugging. The other advantage of the sand filters is they are more reasonably priced.

We discovered a sand loading that results in a high efficiency yet doesn't plug.

Other sand filter media investigated included coarser sand, porous ceramic material, and crushed lava rock. Other hollow media were also looked at. 

While it is true that pressure type filters such as sand filters may require a little more electricity to operate, most pond owners are willing to spend a little electricity to replace their labor. Owners of these filters want something that will do the job better, and with less labor.

In the USA, too often our “Honey-Do Lists” are too long to allow using more labor-intensive filters.

Settling Tank Filters are Labor-Intensive

However, in Japan and China where labor can be cheaper than electricity, the standard for Koi pond filtration has been sedimentation-type tanks in series. They have usually consisted of three to four rectangular or cone-shaped tanks, with outlet valves on the bottom of the cones, and with the tanks plumbed in series.

Peter Waddington in his book "Koi Kichi" describes these systems and some improvements in detail.

The combined volume of the tanks depends on the pond size, and ranges from 6 to 30% of the ponds volume:

Pond size

Flow rate GPH

% of Pond
















The first tank, sometimes referred to as a pre-filter, has its intake coming in tangentially causing the water to swirl in a circular or vortex motion. Large particles of algae and gunk settle down into the bottom of the cone and are drained to waste when the bottom valve is opened.

The second tank has hundreds of bottlebrushes hanging into the water. They provide some further mechanical filtration. 

The third and fourth tanks have layers of a stiff fiberglass like matting material, which do the biofiltration by providing the surface area for the bacteria to live on. Proponents of this system point out the long residence time of the water in contract with the matting materials, which they say allows for better biofiltration. Second they point to the electrical power savings available with these systems.

This system works, but it has 4 drawbacks. First, it is very labor intensive especially when it comes to cleaning out the brushes and matting material. In fact, it is almost a full time, messy and smelly job to maintain. The bacteria in a good biofilter looks like black muck and smells like sewerage. Second, it is not too good for mechanical filtration, especially for very small particles. Third, it turns out to be a paradise for worms. Fourth, it is very expensive with prices of $5,000 being typical.

There is a new centrifuge type mechanical filter that adds even more cost to these filters, but does improve the mechanical filtration. It adds another $2,000 to the total cost.

Cheap Imitations

There are some cheap imitations of this system that use only 1 tank, not the typical 3 or 4, and the 1 tank is also much smaller than 6 to 30% of the pond's size. In fact, for a 5,000 to 10,000 gallon pond, it is not 500 or 3,000 gallons, but the single tank is about 25 to 50 gallons, which is obviously way too small for effective filtration, either mechanical or biological. It is still a messy and smelly job to clean, and it must be cleaned more often than the larger tanks, even though some manufacturers claim they "only have to be cleaned once a year". Some companies charge $500 just to clean them.

These cheap imitations of the Japanese settling tank filters are surprisingly expensive at $1,000 to $3,000, and simply do not do the proper mechanical and/or biofiltration job for a fishpond.

One of the major problems of cleaning the settling tank type of filters is often they are cleaned using chlorinated garden-hose water, which kills much if not all of the biologically active bacteria. This can mean a whole new "ageing or maturing cycle" taking from several weeks to a couple of months before the filter is 100% effective again. This can cause harm to the pond fish suffering through this adjustment with high levels of ammonia and nitrites, which can kill them.

State-of-the-Art Filters

For all of the above reasons the latest generations of sand filters currently represent the “State-of-the-Art” in pond filtration, and they are surprisingly inexpensive, especially compared to the labor-intensive sedimentation tank systems. They have all the benefits of more expensive filtration systems, without the plugging problems, and they are 10 to 20 times more efficient than bead filters.

The most important measure of a filter's efficiency is its cost in dollars per square feet of surface area. In other words, how much does the biological activity you need cost?






Hit Counter

Koi fish, Nishikigoi, Japanese Koi, Show Koi, Koi ponds, Butterfly Koi, Koi diseases, Koi identification pictures and everything on "HOW TO" information!

Koi Koi Sale Prices Other Websites Koi Book Store We Buy Koi Koi Varieties Koi Growth Koi Weight Koi Age Koi History Koi Taxonomy Koi Sexing Koi Breeding Koi Can Hear Koi Can Taste Koi Can Smell Koi Can Blush Koi Can Overeat Koi Nitrification Koi Filtration Pump Savings Pump head Koi Need Oxygen Koi Shows Koi Quality Koi Feeding Why Koi? Koi Imports Koi Pond Liners Koi Pond Tours Koi Pond Plants Koi Predators Koi Patents Koi Equipment Koi Services Koi Store Pictures Koi Water Quality Koi Water Testing Koi Health Koi Pond Heating Keywork Search Koi Index

   Phone: 303-883-8000

e-mail: ColoradoKoi@ymail.com

Postal Address: 5082 E. Hampden Ave, Suite #138  Denver, Colorado  80222

Copyright © 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009,2010, 2011, 2012, 2013, 2014, 2015  Colorado Koi, Inc. All Rights Reserved.