Solids Production & Removal in Intensive Salmonid Aquaculture

By Adrian Desbarats, BSc

The ability to accurately model solids removal is paramount when selecting filtration equipment and when designing a system for maximal solids removal.

However, effective modeling of solids removal efficiencies first requires information regarding total solids output, rearing tank design, mechanical filter performances and intended flow rates at various stages.

With respect to solids output, Merican and Phillips (1985) reported no significant relationship between total suspended solids output and fish size and concluded that feeding rate is primarily responsible for solid waste output.

A number of researchers suggest that a load of 300 gm of solid waste will be added to the system for each 1 Kg of feed (Stechey and Trudell, 1990). Henderson and Bromage (1988) state that suspended solids output as a proportion of feed fed ranges from 0.3 to 0.65 while Babtie International Ltd (1995) suggest that 32% of feed fed results in solid waste.

In each of the above cases, however, there has been no mention of feed characteristics. Over feeding excluded, the primary factor that affects solid waste production is feed digestibility. Every percentage increase or decrease in feed digestibility will result in a corresponding and equal change in solids production. Furthermore, in field determination of solids output is extremely difficult and often inaccurate (Kwan Cho et al, 1999).

Recently, however, Cho and Bureau (1998) have proposed a biological method for determining total solids output from a fish farm. This model assumes that a fish requires a certain amount of energy to meet its basic metabolic requirements and that additional energy beyond this will be used for growth and reproduction. By understanding the metabolic and growth requirements of the species of interest and the nutritional, energy and digestibility characteristics of the feed of choice, very accurate predictions of solids waste production can be made.

Once this biological model was applied, I found total suspended solids output to comprise 25% of total feed fed for a typical Atlantic salmon grow-out operation assuming approximately 5% feed waste.

This figure, although somewhat lower then earlier predictions, is still very comparable. However, it should be noted that this prediction is based directly on feed digestibility which in the case of most Atlantic salmon feeds is around 80%. Because feed digestibility will significantly affect solids output and, without knowing the digestibility of feeds used by other researchers, it is difficult to make direct comparisons. For instance, Cho and Bureau (1998) found that at one field station fed a feed of very high digestibility (approx 85%) solid waste output comprised only 18% of 60t of feed input.

It should be noted that when using a bio-energetic model, estimation of waste output relies heavily on accurate prediction of growth rates. In this case a thermal growth coefficient (TGC) of 0.08 was used to predict growth rates. Although TGC is species specific, it may also vary with site, nutrition and feeding protocols. As such, we have observed TGCs for Atlantic salmon farms anywhere from 0.06 to 0.11. With this in mind, actual solids production will probably range anywhere from 20% to 35%.

At this time, other possible sources of solids should also be considered, these being uneaten feed and biofilter floc. Uneaten feed is the most difficult to determine as it will be site specific. However, a fish farm having good husbandry practices should have an average feed wastage anywhere from 2 - 5%.

Biofilter floc production is directly related to rate of ammonia oxidation. Approximately 0.17 Kg of floc is produced per Kg of ammonia oxidized. For a typical Atlantic salmon smolt operation, fed a diet having approximately 45 - 50% protein, biofilter floc waste will account for only 2 - 3% of total waste production.

Once solid waste output is determined, it is then possible to determine suspended solids concentrations at various points in a system. Armed with this information, it is then possible to evaluate various filtration and design options. Following is a list of common design considerations and filtration technologies:

Design Considerations

The first step in successful solids control must begin BEFORE construction - In other words, good system design. Since faecal matter accounts for > 80% of solids production, rearing tanks should be the first line of defense. Good tank design is critical to effective solids removal - must be able to effectively and efficiently move the solids from the rearing tanks to the filtration equipment. Tank design will be covered in another Tech Talk article, but following are the general principles to follow:

  • Tanks should be circular to produce desired hydraulics. There is some argument regarding tank diameter to depth ratio, however, anywhere from 4 - 10 to 1 will work well.
  • Diffuser design and placement are also critical to create proper hydrodynamics. Ideally diffusers should be in the shape of an inverted L with the lengths of the vertical and horizontal pipes being approximately 30 - 50% of the tank diameter. Tank water velocity will be dictated by the number and size of holes in the diffusers. Velocities of 15 - 30 cm/s are required to drive settleable solids to the tanks center drain. If tank diameter is large (> 10m) two diffusers should be employed (located opposite each other) to ensure adequate velocities are maintained over the entire tank's volume.
  • Dual drain is critical to concentrate solids. Ideally, the majority of the solids should leave the tank bottom drain with only 15 - 20% of the tank flow.

Outside of tank design, the water distribution system should be carefully considered. PVC in-ground pipe should be used versus open concrete channels. In systems I have seen employing open concrete channels, the rough surfaces and lack of adequate flow cause solids to settle out in the channels. Pipe runs must be straight and short as possible to get the solids to the filtration equipment quickly.


This is an older technology but it is still widely used in the aquaculture industry. Sedimentation relies and particle density, removing particles through settling action.


  • Requires no energy to operate;
  • Relatively inexpensive to install;
  • Easily incorporated into existing facilities (providing there is available land).


  • Poor solids capture < 100 um;
  • Low hydraulic loading requirement = large surface area requirement at high flow rates;
  • Can be expensive to maintain - Transport / disposal of sludge (ie. high water content);
  • Maintains solids in contact with water - leads to significant leaching of nutrients such as phosphorous and nitrogen.

Swirl Separation

Unlike hydroclones, swirl separators remove particles through sedimentation NOT centrifugation. Therefore, a well designed swirl separator should consider particle settling velocity, volumetric flow rates and design features that enhance sedimentation. Recent studies have shown that tangential inlets (as are commonly used) result in less efficient solids removal versus non-tangential inlets (Brooks et al, 2001).

Tangential inlets increase water velocity and create turbulence, both of which are counterproductive features in a swirl separator. Inlet designs that show the best removal efficiency consist of a downward spout situated in the centre with a deflection plate. Another efficient inlet design incorporates a rectangular shape.

Regardless of what inlet structure is chosen, the key elements that should be considered for inlet design are reduction of turbulence, slowing and dispersing in-flow and, minimization of short circuiting.

Proper outlet design is also important to effective swirl separator performance. Outlets that offer the best solids removal are weir type designs that encompass the entire periphery of the swirl separator.

Whatever outlet design is chosen, the main concern is to choose a design that slows the water velocity (by having a large volume capacity), creates quiescent conditions and promotes water flow over the entire volume (inhibits short circuiting).


  • Before the development of dual drain rearing tanks and mechanical filtration swirl separators were an effective means of concentrating solids;
  • Require little energy to operate;
  • Easy to operate, maintain and clean;
  • Very efficient if properly designed.


  • Ineffective if improperly designed;
  • Leaching of nitrogen and phosphorous an issue;
  • Poor removal of solids < 150 um;
  • Counter-productive if rearing tanks are properly designed.

Foam Fractionation

Two primary components required for foam fractionation to occur:

  • Presence of a surfactant with polar and non-polar groups;
  • Presence of air bubbles.

As air bubbles are introduced to water, surfactants become hydrophobic - tend to stick their non-polar end out of water (ie. into an air bubble). The following diagram shows surfactant / air bubble interaction:

Air Bubbles

Diagram taken from: Recirculating Aquaculture Systems, MB Timmons et al, 2001

Foam fractionators are most effective at removing fine dissolved and suspended solids - larger solids tend to get removed from air bubbles as a result of excessive shear forces. Foam fractionation will effectively remove particles < 30 um in size.


  • Very effective at removing suspended solids below 30 um;
  • One of only a few technologies capable of removing dissolved organics economically;
  • Oxygenation as an added benefit;
  • Effective at reducing bacteria and viruses through flocculation and ozone neutralization (if ozone is used as a supplement to air addition).


  • Temperamental in freshwater - low PH, presence of oils, alkalinity, temperature fluctuations, water density all play a role in affecting performance;
  • Increasing fractionator diameter leads to reduced removal rates (180 cm maximum effective diameter);
  • At a max flow of 0.15 lpm / cm2 fractionator cross sectional area foam fractionation becomes capital and operating intensive at high flow rates.

Rapid Sand Filtration

Can remove particles down to 10 um. Choice of sand is very critical: grain size, uniformity of size, grain shape and specific gravity all play a role in performance.


  • Minimal space requirements;
  • Cleaning can be easily automated;
  • Very effective at removing fine suspended solids.


  • Aquacultural waste is very "sticky" - can create clogging issues, even with regular back flushing;
  • High head losses (5-20m+) leads to high pumping costs - impractical with high flow rates.

Rapid sand filtration should only be considered for applications having low total suspended solids (TSS) and low turbidity where only "polishing is required. In other words, this technology should NOT be considered for filtration of high TSS high BOD recycle water.


When used on recycle water, ozone is very effective in improving water quality (see the Tech Talk: "Use of Ozone in Intensive Fin-Fish Recirculation Systems"). Ozone is very effective at removing suspended and dissolved solids through oxidation as well as through flocculation.


  • Very effective at removing fine TSS < 60 um;
  • Also improves removal of TSS by mechanical filtration through flocculation;
  • Significantly reduces COD, DOC and colour and reduces levels of bacteria and viruses;
  • Reasonably inexpensive at ~ 10 KWh per Kg O3 produced (assuming pure oxygen feed gas and 4-5% O3 produced).


  • Extremely toxic to fish - However, if applied in conservative doses (13 - 20 gm per Kg feed), retention prior to rearing tanks is 3 min or greater and ORP probes are used, risk is extremely low);
  • High initial capital cost.

Mechanical Filtration

With improvements in this technology, it is very rapidly becoming the filtration technology of choice. Filtration performance is primarily dependent upon screen opening size and influent TSS. Evidence suggests that @ screen openings < 60 um, there is minimal improvement in solids removal.

It has been consistently shown that drum filters and belt filters display improved removal efficiencies with increasing solids concentration (Kwan Cho et al, 1999; Couturier et al, 2001; Summerfelt et al, 2000). In fact, it has been found that efficiency of solids removal is so closely related to solids concentration that regression equations can be applied.

Summerfelt et al (in press) found that capture efficiency for drum filters could be expressed as follows:

Y= 16.229 ln(x) + 32.62 (R2=0.4475)


X= solids concentration (mg/L)
Y=removal efficiency (%)

Roglund Research (in press) found the capture efficiency of belt filters could be expressed as follows:

Y= 21.34 ln(x) + 28.639 (R2=0.831)

For example, by sending the low flow, high solids water from the bottom drain of rearing tanks (assume approximate bottom drain flow to be 20% of total rearing tank flow and solids concentration > 50 mg/L) to a 60 um drum filter, total solids removal will typically range anywhere from 75 - 90%. However, if the total flow from rearing tanks is sent to a group of drum filters (high flow, solids < 50 mg/L) filter efficiency will range anywhere from 40 - 50%.

This example clearly demonstrates the importance of concentrating solids at the rearing tanks prior to mechanical filtration.

Drum Filter - Advantages:

  • Very efficient at high TSS concentrations (> 50 mg / L TSS, average removal efficiency of 90%);
  • Can accommodate very high flow limits, economically;
  • Require minimal space.

Drum Filter - Disadvantages:

  • Less efficient at low TSS concentrations ( < 5 mg / L TSS, average removal efficiency 50% at best);
  • High water content in backwash - 0.03 to 0.05% TSS.

Rotary Disc Filter - Advantages:

  • Lower investment cost versus other types of mechanical filtration.

Rotary Disc Filter - Disadvantages:

  • High backwash volume production;
  • Tend to grind solids;
  • Less efficient at high solids loading.

Belt Filter - Advantages:

  • Removes solids in a gentle manner;
  • Low maintenance;
  • Extremely efficient at low flows and high TSS concentrations ( > 50 mg TSS / L, average > 90% removal);
  • Produces highly concentrated backwash ~ 10% solids.

Belt Filter - Disadvantages:

  • Less adaptable to high flow rates.

Drum filtration in combination with belt filtration is an excellent combination. The drum filter is very effective at capturing solids above 60 um at high flow rates. The drum filter backwash can then be sent to the belt filter for de-watering. In this manner, one belt filter can service the backwash from multiple drum filter units. By employing this combination, sludge production can be reduced by ~ 250 - 300x (reducing sludge handling and transport costs and removing nutrient leaching processes).

Print This Article


Close Window