Print This Page

Water Quality and Feeds

A Discussion on the Shrimp List


This discussion occurred on the Shrimp List (see page 276), under the subject heading of Feed Grades and Ammonia Buildup, from December 3, 2003, to January 16, 2004.  It’s a rambling discussion on water quality and feeds in intensive systems that extends into super-intensive systems.  Some of the information did not appear on the Shrimp List, but was added at a later date for clarification.  Contact information on all the participants appears at the end of the report.


Julio Estrada, a shrimp farming consultant in Ecuador, started the discussion with the following question to the List: Hi All,


What impact should we expect from various protein levels in shrimp feeds on ammonia/nitrite buildup in ponds when raising Penaeus vannamei in low/minimal water exchange systems?


I'm interested in relatively intensive systems, say systems stocked at 50 to 100 postlarvae per square meter with animals larger than four grams.


I understand that a 35% protein feed, with a good amino acid profile, is about right for optimal P. vannamei growth.  Of course, digestibility, cholesterol, the right fatty-acid profile, micronutrients and vitamins have to be there too.  If the percentage of protein is increased, the shrimp will start burning protein for energy, releasing even more nitrogen compounds than usual into the pond.  If a very low protein feed is used, I understand that growth will suffer, but I wonder if protein digestion efficiency (nitrogen conversion efficiency) is known or suspected to suffer as well, again increasing the release of nitrogen into the pond?


To be a bit more specific, has anyone run comparative trials on feeds in the 25%–40% protein range and come up with any sort of correlation between the level of protein and ammonia/nitrite buildup?


Eric De Muylder, a crustacean feeds specialist in Belgium, responds: Dear Julio,


Protein (or amino acids) should be used for growth and not energy.  Therefore, no other nutrients should be limiting, and energy should be available from other sources.  I don't think the protein level in itself would be the determining factor, but more the energy/protein range.  So instead of reducing the protein level, you could also try to increase the energy level.  The easiest way is to increase the lipid level by adding more fish oil (and maybe more lecithin) to the diet.  If you decrease the protein level, and you replace the protein with starch or fiber, which are more poorly utilized by shrimp than proteins and lipids, you are going to increase the organic load to your system, which may be useful to bacteria, but it will indirectly increase the oxygen demand of your system.


Peter Van Wyk, a shrimp culture specialist at Harbor Branch Oceanographic Institution, comments: Dear Eric,


I agree with your statement that, as much as possible, protein (or amino acids) should be used for growth and not energy.  As you pointed out, when feeds have insufficient energy for the amount of protein, protein is utilized as an energy source.  But I am not sure I agree that the best strategy is to limit ammonia buildup by increasing lipid levels.  This might be the case in an algae-based system where solid wastes are filtered out of the system, but, increasingly, shrimp farmers are substituting aeration for water exchange and managing their ponds to stimulate bacteria rather than algae.  Reducing protein and increasing fiber and starch content may work better in these systems.


According to Dr. Yoram Avnimelech, a researcher in Israel who specializes in pond bacteria, the typical protein utilization efficiency for an algae-based system in which the shrimp receive a 35% protein diet is about 20%.  In closed, bacteria-based systems, the typical strategy is to reduce the protein content of the feed, reducing the percentage of high protein ingredients with ingredients that are high in carbohydrates.  Carbon is the primary limiting nutrient for bacteria in typical production systems.  Lowering the protein content of the feeds and increasing the carbohydrate content allows the bacteria to flourish.  The bacteria consume uneaten protein and also utilize ammonia and nitrate as a source of nitrogen for protein synthesis.  If the pond is well aerated, an organic floc develops that is thick with bacteria.  Shrimp consume this floc and convert a good bit of the bacterial protein into shrimp protein.  Shrimp protein utilization efficiencies in well-aerated bacteria-based systems may increase from 20% to nearly 40%.  Properly managed, these systems are characterized by low ammonia concentrations.  Low-protein, high-carbohydrate feeds are cheaper than high-protein, low-carbohydrate feeds, so feed costs are reduced.


When we first started playing with bacteria-based systems here at Harbor Branch (I should note that this was in raceways, not ponds), we initially used a 35% protein diet, and supplemented the carbon levels by adding sugar (not cheap, but it was pure carbohydrate and easy to add).  This definitely put the bacteria in high gear...too much so, in fact.  The sugar was very rapidly metabolized.  Adding sugar we could get ammonia and nitrite levels to drop precipitously over a matter of a few hours.  Unfortunately, DO would also drop precipitously in response to the sugar.  We learned that it was better to add sugar in small amounts over the course of a day, which helped reduce some of the volatility.  We later switched to using molasses because it was cheaper, and we could add it by means of a dosing pump to even out the spikes in oxygen demand.  But even then, we experienced very thick flocs of organic material, very high BOD, high CO2 levels, and low pH.  A big part of the problem was that by feeding a high level protein we had to add an awful lot of molasses to achieve the C:N ratio necessary to control the ammonia and nitrite.  The result was a system with a very high BOD, huge amounts of CO2 production, and low pH.  The system was very difficult to control and unstable.


The key to getting this system under control turned out to be reducing the protein content of the feed.  By reducing the protein content of the feed, less carbon is required to achieve the desired C:N ratio.  This permits ammonia to be controlled without development of such heavy organic flocs.  Bacterial respiration is reduced, reducing oxygen demand and CO2 production.  By using a grain-based, low protein feed (I believe this is Belize Aquaculture's feed strategy), desired C:N ratios can be achieved without molasses, and feed costs are lowered.  Feed conversion ratios may also be lower because the shrimp are deriving a significant portion of the diet from the organic floc.  Another important benefit of using low-protein, grain-based feeds rather than high-protein feeds with molasses is that the grain is metabolized more slowly by the organisms in the system, and this results in a less volatile system.


Julio Estrada: Peter and Group,


Has anybody gotten good nitrogen conversion efficiencies in ponds with soil bottoms and without AquaMats?


I understand the wastewater people use rather extreme aeration and agitation to get proper flocculating, and they take water from an ongoing treatment tank to "seed" the next tank, so getting the flocs started appears nontrivial.


Finally, I have no idea what bacterial floc looks like.  Would someone please describe it to me?


Hank Bauman, manager of Belize Aquaculture, Ltd., probably the most intensive shrimp farm in the world, enters the discussion: Hi Julio,


In our plastic-lined, zero-exchange ponds, we produce from 15,000 to 25,000 pounds per acre per crop with a low protein feed, and ammonia has never been a problem (less than 1 ppm).  But we do get occasional nitrite spikes up to 6 ppm.  Not sure why.  All ponds are treated exactly the same and fed on demand.  Some grow fast and some don't.  It seems to me that we still have a long way to go in controlling the bacteria species, and in keeping out the blue-greens.


Julio Estrada: Hi Hank,


Thanks for the info.  Do you suspect any correlation between nitrite spikes and blue-green blooms (even minor blooms)?  Someone once reported that the blue-green spike comes first, followed by a nitrite spike.


Nitrite is a real problem.  Lots of bugs always around to eat up (or oxidize) ammonia, given enough carbon, but nitrite-eaters apparently won't be there unless you've promoted them by having some level of ammonia always available—plus you’ve got to live right.


Eric De Muylder: Hello Peter,


I agree.  In a bacteria-based system, the whole approach changes.  From my understanding, Julio was not using a bacteria-based system, but a low water exchange system.  In a bacteria-based system, you can lower the amount (not the same as percentage) of proteins you feed the shrimp, and an increased carbon:nitrogen ratio will limit the buildup of ammonia.  I don't think it makes sense economically, however, to increase the carbon in the pond by feeding shrimp starch and feeding the bacteria with uneaten or undigested feed.  There are surely cheaper sources of carbon that can be fed to the bacteria directly.  Instead of reducing the percentage of protein in the feed, you can probably reduce the amount of feed you give to the shrimp.  Less feed and a better FCR will reduce the costs more than using cheaper feeds in most cases.


Eric De Muylder: Dear Hank and Group,


Ammonia is reduced by two kinds of bacteria: those that use the ammonia as a nitrogen source for the buildup of proteins (all of which need a carbon source) and those that convert ammonia into nitrites and relatively harmless nitrates.


Because ammonia-converting bacteria are slow growing, they can only be kept in zero or very low exchange systems.  The only other way they can be kept is if you provide a substrate like a media in a biofilter.  And they only start growing after you have a buildup of ammonia and nitrites.


Phil Boeing, a shrimp farming consultant, adds:


For my two bits, folks, I have been operating commercial recirculation systems for penaeid shrimp for over four years and here is what I know and don't know.


Shrimp culture in very elaborate and well-engineered recirculating systems has not been shown to be commercially profitable due to high capital and operations costs.  Even with a select or niche market, it is difficult to make money.  A simple stacked tray system is probably the best means to get more biomass per volume of water, but it generates difficulties with chores like feeding and cleaning.


Multiphase, recirculating systems, which move shrimp through a series of increasingly larger containers as they mature, show good production numbers, but they are not profitable.


I attempted on several occasions to duplicate the low-protein, high-carbohydrate efforts of Avnimelech and others, but it presented more problems than I already had.  It is very oxygen consumptive because the material is not consumed for some time after it is fed and converted to single cell protein.  It piles up and goes septic...and on and on and on.  A labor nightmare is what I remember most.  Not to say it isn't the way to go, but it does need some careful design and implementation rather than trying to adapt it to an existing facility.


I have grown shrimp for years in zero-exchange systems at 5–15 parts per thousand salinity with no filtration at all.  I found that the ammonia-to-nitrite bacteria are very fussy and difficult to manage.  The issue of nitrite spikes is a huge topic for wastewater treatment engineers.  I mean entire careers have been devoted to this subject and its solutions.  Many nitrite reducers do not like elevated ammonia at all.  They will not tolerate high ammonia, and the fact that ammonia is what is produced first in protein nitrogen reduction makes this a real catch-22, and is one of the causes of nitrite spikes.


With this in mind, I elected some years ago to mass culture select bacteria with specific media and inoculate the recirculating system with them.  The results were unbelievable.  The entire tank and even the shrimp themselves become coated with nitrifying bacteria and the issue of nitrite spike (or ammonia) became mute.  Depending on the density of shrimp cultured, you might have to add some substrate to the system to control the ammonia nitrite levels, but this is far simpler and more cost effective than anything else I have managed.


I am not yet comfortable enough to begin reducing the protein in the feed, but I wish to do this as a controlled effort and see just what happens with densities in the 2–3 kilograms of shrimp per square meter.  The mind-set of the shrimp farmer keeps telling me that oxygen injection is cost prohibitive except for high biomass.  But it may be that the oxygen level is more critical than realized at higher densities.  I have been told by pond farmers that above 1.0 kilogram per square meter, an oxygen level of five parts per million and up will generate 10% faster growth in white shrimp than in equivalent ponds at around four ppm.


Rod McNeil, a shrimp farming consultant, responds:


I'm going to address a number of issues raised in the above discussion:


Transferring shrimp from bacteria-based nursery culture to algae-based growout is practiced widely in southern China.  A continuous bacteria-based system is maintained and animals are transferred into it at PL–8, at a stocking density of 5,000 per cubic meter.  They then remain there until they are sold and price varies by weight.  The two most common sizes are PL-20 (25 mg), at $6.00 a thousand and 1,000 mg at $20 to $24 a thousand.  They are typically sold to co-op farms that rear them to maturity in higher exchange (3-8% a day) algae-based ponds with final harvests at 5-6 tons per hectare per crop.


On the carbon:nitrogen ratio issue as it relates to excess ammonia in bacteria-based ponds, I'll throw my two cents worth in for you all to consider.


As a practical matter, if you have a fully developed microbial floc, your ammonia levels will be stable, probably in the 0.2 to 1.0 ppm range.


I find that the difficult point comes when a pond is making the transition from algae-based to bacteria-based.  Ammonia may reach 5–7 ppm and nitrite may reach 20 ppm during this transition.


I use AquaMats (big surprise) because their surface area allows a stable community of nitrifiers to develop and remain functional.  The most recognized nitrifiers (Nitrosomonas and Nitrobacter) require a stable, three-dimensional colonial buildup on a sessile surface in association with other species to function.  If you don't have much surface area, not much of this type of a community can develop.  This is the reason that it is classically reported that it takes 2–6 weeks for a nitrification filter to become effective.


I agree with the citation related to Dr. Avnimelech's work and have found algae-based, low-exchange systems function at 20–26% total nitrogen uptake efficiencies.  The oxygen demand of a microbial-based system doubles the oxygen requirement over the metabolism of the shrimp themselves and this need, if met, in a floc-dominated, properly managed system, will allow nitrogen assimilation efficiencies (NAE) of greater than 42%.  We use this value as a cutoff to our feed management practices in SIMSS (Super-Intensive Microbial Shrimp Systems).  I must emphasize that this approach only works well with benthic detrital grazers such as P. vannamei.  Trying to apply bacteria-based culture to a species such as P. monodon is a great deal more difficult.  In addition, we have found it very difficult to manage SIMSS (>5 kgs/cubic meter or 50 tons per hectare/cycle) with protein feed values above 22% in growout.  We use nurseries in all our systems.


Julio, I will address some of your concerns: We use AquaMats to provide the needed surface area to remove the ammonia buildup.  Our average production in open ponds, without liners and using PDP aeration, is 12.6 tons per hectare per crop.  This is slightly too low to achieve the full benefits of SIMSS culture techniques because you're sort of stuck in the transition between algal and microbial communities.


We do use AquaMats in some of these systems, and our average production with AquaMats is 5-6 tons per hectare per cycle greater (improved survivals and higher stocking densities) than without AquaMats.  Trials conducted at numerous facilities have shown 7-14% higher survival, at identical stocking densities, when AquaMats are used.  In addition, those facilities stocked at high enough densities to maintain flocs in zero exchange systems (>1.5 kilos/cubic meter) achieve much higher nitrogen assimilation efficiency, typically 42-49% when the proper feeds are administered.  This cuts production costs dramatically.


As for the issue of what protein levels are appropriate, we have standardized on 41% protein feeds from PL-8 to PL-20 and then reduce protein content (as described by MacIntosh) by 1.4% per week for six weeks and then 3% per week until such time as protein levels reach 21%.  We can maintain an average growth of 1.55 grams/week, from one gram, with Shrimp Improvement System’s PLs, and harvest in 110-120 days.


As for what a good floc looks like, I will send you some photos from some of our ponds.  Basically, the floc itself will vary from 0.1–3 mm in size and is very light tan in color and the water supporting the floc has a definite strong tea coloration (don't you just love all this highly detailed scientific description!).  When placed in a beaker or an Imhoff cone, the floc will settle out in less than a minute, reaching full compaction within 15 minutes and typically constitutes 3–5% on a bulk basis.


We find that making iron and silica additions to the pond help in the early formation of the proper flocs (lessons taught to me by Robins McIntosh).


We use a product called "slip," which is generated from bentonite clays and is available from most bentonite suppliers, to provide the 0.3-5 micron silica, which has quite high iron content.  Unfortunately, most of this iron is in the ferrous state with poor solubility, so if we are having problems initiating a floc, we'll add the slip first at 10 parts per million, mix for 24 hours and then add 0.5 ppm ferric chloride.


The flocs stick readily to an AquaMat, where they are grazed on by the shrimp. 


You can certainly get bad flocs, which are almost black in color.  You will have gill fouling and very poor growth once such flocs form.  Another type of floc is bright lime-green and is excellent for growth, but hard to initiate.  Floc management is a definite subjective skill and we manage our flocs by altering the feeding regimen.  If flocs are present with too much foam, we'll stop feeding for 24–48 hours to increase the grazing pressure on the flocs.  This fixes the problems in 90% of the cases.  If we get a black floc established, we'll do a heavy (25%+) exchange and raise the pH with hot lime.  Once the pH is above 8.4, these "bad flocs" don't seem to survive.




Julio Estrada, Lansol, S.A., San Martin 110 y la Ria, Guayaquil, Ecuador (phone 593-4-283-0442, fax 593-9-975-9977, email


Eric De Muylder, VDS Crustocean Feeds, Paanderstraat 40, 8540 Deerlijk, Belgium (phone 32-56-719168, fax 32-56-723002, email, webpage


Peter Van Wyk, Harbor Branch Oceanographic Institution, 5600 U.S. Highway 1, North, Ft. Pierce, FL 34946 USA (phone 772-465-2400, fax 772-466-6590, email


Hank Bauman, Belize Aquaculture, Ltd., #1 King Street, Box 37, Belize City, Belize (phone 501-2-77031, fax 501-2-77062, email


Phil Boeing, Phil Boeing, Shrimp Consultant, 53-300 Avenida Navarro, La Quinta, CA 92253 USA (phone 760-564-1421, email


Rod McNeil, Meridian Aquatic Technology, LLC, 303 Kerr Dam Road, P.O. Box 876, Polson, MT 59860 USA (phone 406-982-3109, fax 406-883-8592, email, webpage


Source: 1. The Shrimp List. Water Quality and Feeds. August 2004. 2. Bob Rosenberry, Shrimp News International. Update. March 22, 2017.



Print This Page