The results of this study, which have been reported in section 4, provide estimates of the possible demand for fishmeal and fish oil if the culture of carnivorous aquatic species continues to expand at similar rates to historical values. These projections are also dependent on the accuracy of the assumptions that have been made on the future levels of commercial aquafeed use, marine resource inclusion, and feed conversion efficiency.
There are many mitigating factors that will influence actual developments. The use of all of the fish oil supply in aquafeeds is unlikely to occur, because of its other uses for direct human food and in the pharmaceutical industry. However, there is no fundamental reason why aquafeeds could not consume the whole of the fishmeal supply. Replacement of the fishmeal currently used in feeds for other livestock is nutritionally easier. Partial or complete substitution already occurs to a limited extent and depends considerably on the relative prices of fishmeal and other animal or vegetable (notably soybean) proteins.
Other factors will come into play before the 100% utilization level is reached, either for fishmeal or fish oil. These not only include purely market or economic considerations (the price-squeeze, or original fishmeal trap) but also other matters which have become prominent in the decade since this topic was first discussed. Some of the mitigating factors that will influence the actual usage of marine resources in aquafeeds are discussed below.
The concept contained in the fishmeal trap is that given existing technology and the apparently limited supply of fishmeal and fish oil, the expansion of some types of aquaculture will, if not brought to a halt, be at least considerably slowed down.
The economists version of this argument is as follows: in the face of stable supplies of raw fish for fishmeal production, the growing demand for fishmeal will drive up the price of fishmeal and fish oil to such a level that fish and shrimp farmers will not be able to afford to buy aquafeeds that contain adequate amounts of these marine resources.
A closer look at this reasoning reveals the following assertions/hypotheses:
1. Demand for food fish will grow (FAO, 2000).
2. Given the overexploitation of wild fish stocks, increased supplies of food fish can in the long term only come from aquaculture (FAO, 2000).
3. The proportion of aquaculture that relies on feeds with fishmeal and fish oil ingredients will grow, and will do so rapidly (as discussed in this paper).
4. The growing aquaculture production will need an increasing share of fishmeal to be converted into aquafeeds (discussed in this paper).
5. However, fishmeal production has been static during the past decade and is likely to remain so.
6. Therefore the price of fishmeal will increase drastically and aquaculture must gradually reduce its reliance on fishmeal as a fish/shrimp feed ingredient during its further production expansion.
A closer look at affirmations 5 and 6 is merited. However, before doing so, it is useful to recall that fish and shrimp farmers demand for fishmeal and fish oil is a derived demand. It is the price that the consumer is prepared to pay for cultured fish and/or shrimp that determines what fish and shrimp farmers can afford to pay for the various services and inputs that are required in their production process; amongst these inputs aquafeeds (with fishmeal/fish oil ingredients) are some of the most important.
The sequence through which prices are determined can be described as follows:
1. Consumers determine the upper price level for farmed fish and shrimp.
2. Fish and/or shrimp farmers determine the upper price level they are willing to pay for aquafeeds (and other supplies and services).
3. Aquafeed manufacturers determine (normally using least cost formulae) the maximum price they are willing to pay for aquafeed ingredients, including fishmeal and fish oil.
4. Fishmeal and fish oil manufacturers determine the price levels they are willing to pay for the various fish species and fish offals available as raw material.
5. Fishermen determine the price levels for raw fish at which they are willing to fish for the fishmeal species.
These price levels fluctuate in the short term, following the changing levels of supply and demand in the various markets. In the long term they are influenced by technological developments and, of course, by the bargaining power of buyers and sellers. In addition to the importance of events within the fishing industry, developments in the livestock sector will also influence the use and availability of fishmeal and oil.
The important feature of the above relationship is that in the short term - in which technology is almost fixed - it is the consumer who decides the maximum price levels throughout the chain. In the long term technologies will be modified. This will cause fish and/or shrimp farmers to use a different combination of goods and services which, in turn, will affect the composition of the costs and the prices that they can afford. Examples of the visible economic effect of technological modifications are: improvement in feed conversion ratios; decreased inclusion rates; etc. The current paper has already discussed such technological changes but has not, so far, considered technological developments in the animal feedstuff industry in general, or in the fish oil and fishmeal industry, or in the fishing industry itself.
Technological developments in the animal feedstuff and fishmeal/fish oil industries will not be discussed in this paper - on the assumption that technology is advanced and cost reducing modifications are likely to be minor in the next decades. However, some technological modifications in fishing and the implications of those in economic terms are discussed below.
A number of the following questions are discussed in this section:
Firstly, is fishmeal production likely to remain stagnant?
Secondly, will higher prices for raw fish attract greater supplies of raw material for the industry?
Thirdly, if this does not turn out to be the case, and raw material supplies in fact remain essentially stable, would the price levels for raw fish become so high that, when they have worked themselves through the industry, the resulting aquaculture products would be so costly that consumers will not buy them?
Fourthly, before tackling these earlier questions, it is necessary to look at another threat to supplies: will the growth in demand for food fish mean that some of the fish now converted into fishmeal will be sold directly for human consumption?
Lets look at this last question first:
will shortages of food fish lead to less raw fish being available for fishmeal production?
At present about two thirds of the worlds raw material for fishmeal production is the result of dedicated fisheries. In such fisheries one or two species generally account for as much as 90% of the supplies. In most fisheries the target species are small pelagics, but also some demersal fish of small size are exploited. Dedicated fishmeal fisheries are mostly undertaken by specialized vessels that fish specifically for the fishmeal industry. Large dedicated fisheries are found in Chile, Denmark, Iceland, Peru and the USA. Although Japan is a relatively large fishmeal producer it does not have any major fishery dedicated to supply Japanese fishmeal plants with raw material.
With very few exceptions, the species caught by the specialized fleets have not been and are not being used for direct human consumption, except in small quantities. The main reason is not price. The catch that is used for human consumption generally fetches a higher price than the catch used by the fishmeal industry. The non-fishmeal markets - that is fish for human consumption in one form or another - are small. In fact, the species used for fishmeal processing often present characteristics that make them less than ideal for direct human consumption: the fish are usually small (which means they are difficult to process mechanically), difficult to maintain in good condition once out of the water (Teutscher, FAO, 2001b), and available in very large quantities - sometimes in very sparsely populated areas (e.g. Iceland, the northern part of Chile, Peru). An example of the disposition of catch in Chile is given in Table 6, while Table 7 shows the composition of the raw material used by the Danish fishmeal industry.
Demand for food fish in wealthy countries will increase only modestly in volume terms but shift from low value to high value products. Populations in Africa and South Asia which might be able and willing to purchase some of the fishmeal species for direct human consumption at their present ex-vessel price levels, would find it difficult to do so. The ex-vessel price would be significantly increased to cover preservation and transportation costs from the distant fishing grounds in South America to markets in other continents.
It is therefore unlikely that the worlds fish consumers will dramatically increase their demand for fishmeal species. However, this does not improve matters much for fishmeal producers as the species that now are targeted are close to fully exploited and do not in reality offer much hope for any consistent increase in supply, even at significantly higher prices.
The next question is:
will higher prices for fishmeal - and therefore for raw fish supplies - lead to new dedicated fisheries?
The hope for increased fishmeal and oil supplies lies in the use of species that hitherto have not been used for fishmeal production. The two main sources are mesopelagic species and krill. Both species have been caught and used to produce high protein meals. The problem to date is a techno-economic one: with present fishing technologies, the harvesting, preservation and processing costs are in excess of those that fishmeal producers are prepared to pay.
The primary issue here is:
what prices do fishmeal producers pay for raw fish today?
Table 8 provides some recent figures, which clearly show that prices for raw material differ from species to species and fluctuate from year to year. Prices also fluctuate within each season. In some fisheries market prices do not exist because the fleet is owned by the processing industry. This seems to be the case in parts of South America and in the USA (menhaden fisheries). In Denmark the vessel owners own the processing plants and there is a tendency to pay as high prices as possible for raw material.
However, as discussed in section 5.1.1 of this paper, the price that the fishmeal producer is prepared to pay is directly linked to technological and economic developments.
The next question therefore is:
if the price to the fishing vessel was doubled, tripled, or quadrupled over a 5 to 10 year period - as a result of increased real prices for the final shrimp or fish product, or because of technological developments - what would then happen to the volumes of fish supplied to the fishmeal industry?
To consider this question it is necessary firstly to establish what prices are at present paid to the vessels supplying raw fish to fishmeal plants. Table 8 indicates that the ex-vessel price should be somewhere in the range of US$ 80 to US$ 120/tonne for sustained commercial fishing to be possible.
are there alternate sources of raw material and what would the effect of their use on costs be?
Since the 1970s FAO has been involved off and on in activities aimed to catch mesopelagic species for the purpose of producing, primarily, fishmeal. These activities have taken place particularly in the North Western part of the Indian Ocean. Over a similar period a number of long distance fishing nations (e.g. Poland, Japan) have been fishing for krill in the Antarctic, concentrating on the areas south and east of Argentina. Although krill products for human consumption have been produced, present efforts (by the Japanese) seem to be oriented primarily towards making animal feeds.
Mesopelagic species are defined as species spending the day at depths between 200 and 1 000 m; generally they migrate to 200 m, and at times to the surface, during night-time. Mesopelagic species are found in all oceans but the number of species and, in general, the annual production are highest in subtropical and tropical seas (Gjosaeter and Kawaguchi, FAO, 1980). Experimental fishing has been carried out, particularly in the Gulf of Oman. Catch rates as high as 30 tonnes/hour (Thiele and Valdemarsen, FAO, 2001c) have been recorded and fishmeal of an acceptable quality has been produced. However, commercial production has not been achieved. It appears that specialized vessels are needed. They need to be able to handle the catch in bulk. Also, it is most probable that the attempt to produce fishmeal on board, instead of on shore - which is the norm - proved too costly. The Japanese fishing industry has also been investigating mesopelagic species.
Japan has also been exploiting krill in Antarctica for some time. Most of the krill caught has been used in animal feeds, including conversion into krill meal. The bottleneck at present is that the ex-vessel price for krill meal is about twice that of normal fishmeal (B. Yoshitomi, pers. comm., 2001).
As time goes by, it seems likely that catching and processing technology will improve for krill and mesopelagic species, and that the real price of fishmeal will increase. At some point fishmeal production using these raw materials will become economically feasible. As the stocks of both krill and mesopelagics are large the real price of fishmeal will probably stabilize once these species are exploited by the fishmeal industry.
The next two related questions are:
at what fishmeal price will dedicated fishing on krill and/or mesopelagic species become profitable? and would it be enough if the average price for fishmeal doubled (i.e. reached about US$ 1 100/tonne) and fish oil prices followed a similar pattern?
As the cost of raw fish used in the manufacture of fishmeal currently is equivalent to about two thirds of the international price of fishmeal (which, for the purposes of this study has been derived from the data in Table 9 for the years 1997-1999 as about US$ 550/tonne) it would mean roughly that - other costs unchanged - the raw fish price could increase from an average of US$ 92/tonne to US$ 230/tonne, if the international price for fishmeal doubled. However, this is an extreme assumption for two reasons. On the one hand the cost of other services and goods may increase as well, leaving the fishmeal manufacturer with no possibility to pay as much as US$ 230/tonne of raw fish. However, on the other hand, it is likely that both fishing technology and fish processing technology for mesopelagic species will improve. Given the past fishing experience - at least for mesopelagics - fisheries for these species may become profitable even before fishmeal manufacturers are prepared to offer US$ 230/tonne for raw fish.
The composition of krill is unlike that of most fish used for fishmeal. Krill contains, in relative terms, large amounts of fluorine. Thus a straightforward exchange from fish to krill would not be possible for the manufacturer. Feed manufacturers would need to modify their feed formulations if they used krill. Similarly, the oil from krill has different characteristics to the oil from fish, and the extraction rate of oil from raw krill would be different.
The issue now becomes:
what would the effect on the farm gate price of cultured shrimp and/or fish be if the fishmeal price in fact increases to US$ 1 100/tonne?
This question is discussed in the following section of this paper.
Salmon and shrimp aquaculture are two of the most intensive users of fishmeal.
The Norwegian salmon culture industry (Fiskeridirektoratet, 2001) reports a feed conversion ratio (FCR) of 1.22:1 (one kg of salmon from 1.22 kg of fish feed) for the year 2000. Using a fishmeal inclusion ratio of 40% for that year (see Annex 1, Table 3), it can be calculated that each kilogram of salmon has fed on 0.488 kg of fishmeal. Thus, at a fishmeal cost of US$ 550/tonne, the fishmeal in the diet cost about US$ 0.27/kg of salmon produced. If the price of fishmeal were to double, from US$ 550/tonne to US$ 1 100/tonne, each kg of fishmeal would become US$ 0.55 more expensive, and the production costs for the salmon farmer would increase by about US$ 0.27/kg of salmon produced. The cost of producing a kilogram of salmon in Norway was stated to be US$ 2.17 in the year 2000 (Fiskeridirektoratet, 2001). If fishmeal prices in 2001 had been double what they were in 2000, and all other costs had remained the same, it would have had the effect of increasing the salmon farmers production costs by 12.4%.
However, increases in the real price of fishmeal will be gradual. It will be several years before the price will have doubled. By that time the technology will have improved, in the sense that feed conversion rates will have improved and inclusion rates been lowered. This study foresees that the FCR will be closer to 1.0 by 2015 and the inclusion rate have come down to 25%. On these assumptions, at a fishmeal price of US$ 1 100/tonne, the cost of the fishmeal component of the fish feed needed to produce one kilogram of salmon would be at about the same level as has been calculated for the year 2000, namely US$ 0.27/kg of salmon produced - in spite of the fact that the price for fishmeal will have doubled. However, though predicted improvements in FCR and decreases in the inclusion rate of fishmeal indicate that the cost of the fishmeal component of feeding costs is unlikely to rise by 2015, the evolution of total feeding costs is difficult to forecast. The major factor will be the cost of the fishmeal replacement ingredient(s). If the replacement ingredient(s) are cheaper than the (2000) cost of fishmeal, feeding costs per kilogram of salmon produced will fall. However, if the replacement ingredient(s) in 2015 cost the same as (or more than) the cost of fishmeal in 2000 that has been used in this study (namely US$ 550/tonne) and the cost of fishmeal doubled total salmon feeding costs would increase, despite improvements in FCR and reductions in fishmeal inclusion rates.
Similar reasoning can be applied to shrimp culture - and to the culture of any species that need fishmeal in the diet. For example, this study indicates that the FCR achieved in commercial marine shrimp culture in 1999 was 1.80. Using a fishmeal inclusion ratio of 25% (see Annex 1, Table 3), this means that each kilogram of shrimp produced has been fed 0.45 kg of fishmeal. In todays prices this means that the fishmeal contribution to the production cost was about US$ 0.25 per kilogram of shrimp produced. The farm gate value of shrimp is significantly higher than it is for salmon, reaching US$ 6 - 8 (Hishamunda and Manning, FAO, in press). Thus the cost of the fishmeal used in producing each kilogram of shrimp is equivalent to less than 5% of the farm gate price. If fishmeal prices were to double today - and aquafeed manufacturers passed all of the increase on to buyers of fish and shrimp feed -the cost of producing cultured shrimp need not increase by an amount that is larger than 5% of the farm gate price.
Environmental, social, and ethical discussions about the expansion of aquaculture have been prominent for a long time, and have resulted in a number of codes of conduct and strategies (e.g. FAO, 1995; Svennevig, Reinertsen and New, 1999; NACA/FAO, 2000) and codes of practice and certification, including those of the Global Aquaculture Alliance and Scottish Quality Salmon (Global Aquaculture Advocate, 2001). As part of these developments, concerns about the continued use of marine resources have been expressed. Such considerations were crystallized by an article in the journal Nature (Naylor et al., 2000). These authors argued, inter alia, that aquaculture must reduce the inclusion of marine resources in aquafeeds because the culture of carnivorous species was believed to be contributing to the over-exploitation of certain types of fisheries, with concomitant effects on the stocks of other wild fish. This paper claimed that ever-increasing amounts of small pelagic fish would be caught to supply the aquaculture industry and that the appropriation of aquatic productivity for aquafeeds reduces supplies of wild fish that could potentially be directly consumed by humans. It is unfortunate, in the light of more recent developments related to the bovine spongiform encephalopathy (BSE) crisis (see below), that one of the suggestions made by Naylor et al. (2000) was that the problems caused by inappropriate amino acid balance and poor protein digestibility (which apply when vegetable proteins are used to replace fishmeal) could be partially overcome by the inclusion of meat by-products. The paper by Naylor et al. (2000) generated considerable press comment and many NGOs have alerted the public to the issues concerned. However, in this context it is worth noting that there has been no upward trend in the catch of fish for feed in the past twenty years (see Figures 1 and 2), and that the alternative use of the species utilized for fishmeal and fish oil production for human food have so far proved uneconomic.
Positive arguments in favour of the utilization of wild fish as sources of feed for farmed fish have also been put forward. Åsgård and Austreng (1995) compared the relative efficiencies of captured and farmed fish. These authors calculated that 10 kg of capelin (one of the species caught for processing into fishmeal) could produce 4.6 kg of farmed salmon, of which 3.0 kg is edible. On the other hand, the same amount of capelin would produce only 2.0 kg of wild cod, of which a mere 0.7 kg is edible. This topic was further developed by Åsgård et al. (1999).
Other more general and wide-ranging attacks on aquaculture have been made, which particularly target two major users of marine resources in aquafeeds, namely salmon and marine shrimp farming. Some of these assaults on the industry are widely dispersed through the internet, and have generated a lot of media attention (e.g. Dowden, 2001a,b; Leake, 2001; Girling, 2001). If this type of publicity succeeds in reducing consumer readiness to purchase aquaculture products, it would obviously also affect the requirements for marine resources for aquafeeds. However, public aquaculture organizations, as well as the aquaculture industry itself, are becoming increasingly alerted to the difficulties being faced by those seeking to increase the supply of fish through aquaculture.
The environmental and ethical issues currently being raised by NGOs and the media are important and may have a considerable influence on the actual utilization of marine resources in aquafeeds. The possible long-term effect of public exposure to this issue remains to be seen. However, it is a factor that must be taken into consideration when assessing the forecasts made in this study.
Fishmeal is the most appropriate and (amino acid) balanced protein source for aquafeeds for carnivores. It also appears to contain unidentified growth factors and is an attractant. Even if fishmeal is partially or completely replaced by other protein sources, other products from the industry, such as hydrolysates and attractants, are likely to continue to be used.
Potential pressures on fishmeal and fish oil supplies have been the main incentive for research into means of wholly or partially replacing them in animal feeds, particularly in aquafeeds. Recently, the fishmeal and fish oil industry has had to face other real or perceived problems caused by general concerns about the relationship between the quality and composition of animal feeds and animal and human health.
Traditionally, in common with feed ingredients from other animal sources (mammalian, poultry), the raw materials used in the production of aquafeeds have been regarded as potential sources of agricultural chemical residues, microbial pathogens and heavy metals. After processing, they can also become sources of mycotoxins and microbial pathogens. These topics, together with the problems connected with transmissible spongiform encephalopathy (TSE), of which the bovine form (BSE) is an example, were discussed at an FAO consultation on animal feeding and food safety in 1997 (FAO/ESN, 1999).
Two potential problems have become particularly important recently. The first is the presence of dioxin and PCB residues in human food products of animal origin and the potential carry-over of these substances from animal feeds. The second is the relationship between meat and bone meal and the incidence of bovine spongiform encephalopathy (BSE) in ruminants, coupled with the linkage with Creutzfeld Jacob Disease (CJD).
There is no compelling evidence that farmed fish contain generally higher dioxin residues than wild fish. In a study of European fish cited by Klinkhard (2001), one of the highest dioxin contents found in samples taken between 1995 and 1999 was in wild salmon from the Baltic (Sweden). Of the farmed salmon and trout analyzed during this period from Finland, Germany, Norway, Sweden and the UK, the highest level of dioxin reported was only 15% of the level found in Baltic wild salmon. However, fishmeals and fish oils of European origin have been reported by the Scientific Committee on Animal Nutrition (SCAN) of the European Commission to contain much higher levels of dioxin than those originating from less industrialized regions such as the waters off Peru and Chile (SCAN, 2000). Such differences in dioxin content not only affect fishmeals and fish oils but also influence the residue levels in wild fish caught for direct human consumption.
The EU is proposing that maximum levels of dioxins in fish, fishmeal, fish oil and aquafeeds should be set for the period 2002-2005. The proposed levels are close to the medium levels found in fishmeal and fish oil of European origin but much higher even than the highest levels found in products originating from Chile and Peru (Annex 3).
The comparisons between different sources of fishmeal and fish oil involve very low levels of dioxin. SCAN commented that no adverse effects from dioxins would be expected in mammals, birds and fishes exposed to the current levels of background pollution (SCAN, 2000). Despite this, a considerable proportion of the population of Europe (and undoubtedly other regions) is exceeding the tolerable weekly intake (TWI) levels for dioxins set by various authorities. As there is a considerable safety factor imposed on TWI, this does not necessarily mean that there is an appreciable risk to individual health. However, exceeding TWI levels erodes the protection of this safety factor. Food contributes more than 80% of our daily dioxin intake.
Two further factors are relevant when considering the impact of dioxin residues in the context of this study. The first (and favourable) factor is that our exposure to dioxins and PCBs is decreasing (by a factor of about 50% over the past 10-15 years). The second is that there are other major food group sources of exposure to dioxin besides fish and fish products. Obviously, the amount of dioxins to which humans are exposed depends on the nature of their diet and per capita consumption. In the German study cited by Klinkhard (2001), milk and milk products contributed 39% of food contamination with dioxins. Meat and meat products and eggs and products with egg contributed another 30% and 11% respectively. Fish and fish products contributed only 11%. It may therefore be ineffective to target only fish, because fish is not the sole (or even a major) source of dietary dioxin intake. Furthermore, as indicated above, the dioxin levels in fish vary according to the origin of the fish and their diet. Having recognized this, however, it is still incumbent upon both feed manufacturers and aquaculture producers to be extremely careful about the sources of fishmeal and fish oil used in aquafeeds. It would also be wise for those who directly consume fish oils for pharmaceutical purposes to consider their origin.
Concerns about the levels of dioxin in food obscure the real problem: the sources of contamination - metallurgical processing, bleaching processes in paper production and dry cleaning, other manufacturing processes, and combustion (waste incineration and domestic heating). Our food, including the products of fisheries and aquaculture, may expose us to harmful substances, but food is not the ultimate culprit.
Transmissible spongiform encephalopathy (TSE)
First of all, it is important to state that there is no epidemiological evidence for the transmission to humans of a variant of CJD caused by prions that use fish or fish products as vectors (Globefish, 2001a).
A temporary EU ban on the use of animal proteins in certain livestock feeds was approved at an emergency meeting of the European Farm Ministers on 4 December 2000. This ban has since been extended. The main purpose of the action by the EU was the removal of meat and bone meal from European animal feeds, together with the destruction of stocks of this material, in an effort to contain the spread of BSE. The news (Chamberlain, 2000) that recent research has shown that it is feasible to use meat and bone meal to replace 100% of the fishmeal in marine shrimp diets without depressing performance has come at a particularly unfortunate time; it may be difficult to apply such research results in a climate where any animal fed with meat and bone meal may be regarded by the public as tainted.
The EU ban on the use of animal proteins includes the use of fishmeal in ruminant feeds but does not ban its use in feeds for pigs or poultry, or its use in aquafeeds. The ban on the use of fishmeal in ruminant feeds was initiated because meat and bone meal has unfortunately been used at times to adulterate fishmeal, in order to alter its protein content. IFFO regards the EU ban as being non-scientific, and possibly a form of trade barrier or political move (Millar, 2001). It is probable that only a reliable and simple means of differentiating between terrestrial and marine animal proteins will solve this problem.
While the use of fishmeal is not banned in feeds for other animals, including fish, the ban concerning ruminant feeds causes a further problem for feeds manufacturers generally. This problem is that cross-contamination may occur between batches of feeds made for one type of livestock and batches made for other types of animals. The need to demonstrate that no fishmeal has entered ruminant feeds by carry-over from fish (or pig or poultry) diets makes it necessary to accelerate progress towards dedicated feed mills, which manufacture feeds for one type of animal only.
If past experience can be relied upon, the inclusion of fishmeal in any ban will affect public attitudes towards the use of marine resources in aquafeeds. The public will question the wisdom of feeding fish to fish as well as the sense of feeding animals to animals. These scenarios are not equivalent. On the one hand, ruminants are herbivores and it could therefore be claimed that feeding them animal products is unnatural. On the other hand, the fish species that are fed aquafeeds containing marine ingredients are carnivores, so nothing unnatural is occurring. Despite these facts, public concern exists. Similar concern will be generated by reports about the levels of dioxin in animal products, including fish oil and fishmeal. These concerns will tend to exacerbate the public image problems already identified in section 5.2.
The public image of farmed fish and crustaceans fed with fish and crustacean by-products is likely to be affected, whether such concerns are based on real or imaginary threats to human health. Dioxins are present in both and wild and farmed aquatic products. There is no evidence that any TSE has been transmitted to fish by the use of fishmeals (let alone any link to human disease). People may perceive further, but unsubstantiated differences between wild and farmed aquatic products, This may not only affect the aquaculture industry in the developed countries (e.g. in Europe, North America, Japan and Oceania) but also those developing countries which export high-value aquaculture products to these locations.
Unless such public attitudes can be avoided (through proof of safety and the provision of balanced information), serious constraints on the use of marine resources in aquafeeds may be imposed.
Many plant and animal proteins have some potential as fishmeal replacers. Tacon (FAO, 1994) listed a large number of possible fishmeal replacers, including invertebrate animal by-products (e.g. silkworm pupae, earthworms, zooplankton), vertebrate animal by-products (e.g. blood meal, liver meal, meat and bone meal, poultry by-products), single-cell proteins (mainly from fungal and bacterial sources), oilseeds (e.g. soybean, rapeseed, sunflower, cottonseed), legumes (e.g. beans, peas, lupins) and miscellaneous plant protein products (e.g. corn gluten meal and concentrates made from potatoes and leaves). The major constraints identified by Tacon (FAO, 1994) were:
Limited availability and cost for single-cell proteins.
Lack of palatability and anti-nutritional factors in poorly processed plant oilseeds and legumes.
Limited availability, erratic quality and microbial contamination in terrestrial animal by-product meals.
Palatability problems, and limited availability and high cost for miscellaneous plant proteins.
According to New (2001), generally poorer digestibility, lower availability of some essential amino acids, palatability problems, and, in some cases, the presence of anti-nutritional factors, have limited the replacement of fishmeal by plant proteins. To some extent these factors have been ameliorated by the inclusion of supplemental (synthetic) amino acids and flavour enhancers. More recently, the use of enzymes to enhance the nutritional value of diets based on plant proteins has been suggested and, according to Gérin (1999) used on a confidential basis in aquafeeds. New strains of plants, with lower levels of phytates and anti-nutritional factors may also be developed. Furthermore, if public opinion allows, plants may be genetically modified to improve their fatty acid and amino acid profiles (Chamberlain, 2000). Other alternative protein sources, such as single cell proteins (SPC), have also been considered (Tacon, 1995; Åsgård et al., 1999) but few are available in commercially sufficient quantities, or at prices that would make them serious contenders for inclusion in aquafeeds at this moment. Brandsen, Carter and Nowak (2001) cautioned that fish growth is not the only factor to be considered when assessing potential replacement ingredients for fishmeal, saying that the effects that these may have on disease resistance and immune function needs investigation but is seldom mentioned. Such effects may be beneficial or detrimental.
Feedstuff manufacturers have a natural resistance to replacing marine resources in their products until they are convinced that the good performance that they have achieved in aquaculture production through their use can be replicated by any alternative ingredients. Although research continues to be conducted into the replacement of fishmeal in aquafeeds for certain species, the application of such knowledge will not necessarily be immediately applied. Feed manufacturers tend to be cautious and conservative. This is understandable, since the natural reaction of any farmer who has (say) a disease problem, or whose stock do not perform so well as previously, or whose products become the subject of consumer criticism, is to blame the supplier of feeds first.
Real or perceived dangers in the use of other high protein ingredients may cause aquafeed manufacturers to wonder where alternative proteins can be identified in the current climate. The use of certain animal proteins is already suspect and some plant proteins are being criticized for being contaminated with GMOs. According to Martín (2000), feed companies in Europe are already tending to avoid the use of terrestrial animal proteins (because of the BSE crisis) and genetically modified plant-based ingredients (because of fears of unknown effects) in order to promote consumer confidence.
Theoretically, a ban on (or a deterrent to) the use of any high-protein ingredients creates a large potential market for replacements. For this reason, Globefish (2001a) thinks that the ban on meat and bone meal may dramatically increase the demand for fishmeal, and therefore its price. If this proved to be true, the aquaculture industry would experience enhanced competition for its supplies of fishmeal in the future. However, the fishmeal industry itself remains worried about the damage to the image of fishmeal that has been caused by its association with other animal proteins and the unfounded but inevitable linkage to the BSE/CJD fears of the consuming public (FIN, 2001). The new President of IFFO has forecast reduced EU imports of fishmeal in 2001, compared to 2000 (Millar, 2001).
The study shows that aquaculture has the theoretical capacity to totally absorb global supplies of fish oil, unless current inclusion rates decrease more rapidly and/or the characteristics and rate of expansion of the rearing of carnivorous species envisaged in this study is not fulfilled.
Although fish oils are sources of the fatty acids that are essential components of aquafeeds, there are (at least) partial alternatives. For example, Rosenlund et al. (2001) have shown that replacing up to 50% of the fish oil in high-energy salmon diets with rapeseed, linseed, poultry, palm or soybean oils had no significant effect on growth, survival, or body traits. However, these authors found that the use of fish oil substitutes did have a marked effect on the fatty acid profiles of the farmed salmon and, in some cases, an impact on the lipid content of the salmon fillets. Currently, urgent research on the feeding of lipids to farmed fish (specifically salmon, trout, sea bass and sea bream), including the RAFOA programme, is being conducted in academic and commercial laboratories in several European countries (e.g. Stirling, 1999-2000; Cailliez, 2001). Clearly, as in the case of fishmeal substitution, factors other than the growth and survival rates of the species farmed need to be taken into careful consideration as fish oils become, at least partially, replaced with other lipids. These include possible changes in fillet and processing quality, the sensory characteristics of the farmed salmon, its nutritional value (to humans), and product safety.
Some partial substitution is already occurring in commercial aquafeeds for some species groups, notably salmon. Fish oil usage reduction in this and some other groups has been taken into account in deriving the projected inclusion rates for 2015 and 2030. On the other hand, there is potential for the level of fish oil in aquafeeds for other species groups, notably crustacea, to be increased.
A proportion of the current fish oil inclusion can certainly be substituted from other sources. Many of the products listed above as potential fishmeal replacers are also partial fish oil replacers. However, balancing the fatty acid composition of the diet is not simple when using plant sources and may be resisted by feed manufacturers until supply and economic forces dictate. Substitution is governed by several important factors:
(1) changing the fatty acid profile of the feed immediately affects the composition of the farmed products. There are already differences between the fatty acid composition of farmed, compared to wild fish, which could be exacerbated if fish oil inclusion rates are reduced too far.
(2) the total lipid content in some aquafeeds (e.g. salmon) has increased markedly in the last decade because lipids provide a relatively cheap source of dietary energy. This also affects the composition of the farmed product, compared to wild fish.
(3) the use of alternative sources of dietary lipids, whether from animal or vegetable sources, may affect the taste of the product. The rejection of some consignments of Norwegian salmon in the Japanese market has been reported, for example, following complaints about their vegetable taste.
(4) there will be an increasing effort to reduce feeding costs, since the income earned by aquaculture producers per kilogram produced tends to decline as products become more widely available and often, therefore, cheaper (e.g. salmon, seabass and seabream). Thus, fish oil replacement will be partially subject to prevailing costs of alternative lipid sources.
In summary, marked changes in the lipid composition of feeds for carnivorous aquatic species are inevitable in the future. These will be dictated by supply and economic factors and may affect both the source and the total inclusion rates of lipids. Further research is necessary to ensure that the quality and consumer acceptability of the farmed products remain acceptably high as these dietary modifications evolve.
Other marine ingredients
It is not surprising that fishmeal and fish oil play such a pivotal role in aquafeeds, particularly those designed for carnivorous species, because fish (together with other aquatic animals) form part of the natural diet of wild aquatic animals. Feeding trials have repeatedly demonstrated that, on a purely nutritional basis, the best food or feed ingredient (in terms of palatability, growth and food conversion efficiency) to feed carnivorous aquatic species is another fish or fish product, such as trash fish or fishmeal (Tacon, FAO, 1993b). However, the aquafeed industry has long been alerted to the need to conserve the apparently finite supplies of conventional sources of marine ingredients, principally species caught specifically for reduction into meals and oils but also fish processing by-products. Other marine ingredients, such as fish protein hydrolysates, fish silages, and squid liver meal and squid oil. Supplies of squid meal and oil are scarce and expensive, and their use can only be justified in very small quantities, mainly as attractants. Fish silages have some palatability problems and are generally bulky to store.
Krill is potentially an excellent nutrient source for feeding farmed fish and crustaceans. Besides providing protein, energy and palatability, it is also a source of essential amino acids, fatty acids and other nutrients. In addition, it has the potential to enhance the pigmentation of aquaculture products, thus increasing their visual quality. While it is estimated that the available stock of krill (Euphausia superba) exceeds 35 million tonnes annually, only about 80 000 tonnes/year is actually caught (Yoshitomi, 2001). Increased usage of krill resources, either through direct feeding to cultured fish and crustaceans or reduced to meals for use in compound aquafeeds, has undoubted theoretical potential. However, its actual inclusion as a standard major aquafeed ingredient depends on the ultimate costs of fishing and processing krill, and of transporting it from the catching areas to the locations where the farming of carnivorous fish and crustaceans, and the production of aquafeeds, occurs.
Similar considerations apply to the possibility of exploiting currently underutilized fisheries for the fishmeal and fish oil industry. These resources include deep-sea fish, whose exploitation was discussed by Noguchi (2001). Potential for use of some of the fish discarded by the fishing industry may also exist. Other comments on the potential use of krill and mesopelagics are contained in section 5.1.2 of this paper. Increased utilization of fish and crustacean processing wastes, including those from aquaculture, is also possible, although there are some potential (animal) health hazards that would require attention.
Research on the suitability of all unconventional marine resources as ingredients will be necessary before they become fully acceptable to the aquafeed industry. Ultimately, assuming nutritional quality and safety is assured, the use of all these potential sources will depend on economic factors (see section 5.1 of this paper).
As shown in section 4 of this paper, aquaculture has the capacity to totally utilize all supplies of fish oil and fishmeal within the period covered by this study, if no extraneous factors constrained this occurrence. However, aquaculturists will have to compete for these finite resources in the market place. To the extent that they are successful they will de facto need to demonstrate that the use of fishmeal in aquafeeds is efficient and sustainable, from economic, nutritional and environmental points of view. This aspect was briefly mentioned when feed conversion efficiency was discussed in section 3.3 of this paper. The concerns of Forster and Hardy (2001) that proper means of measuring and recording the relative efficiency of aquaculture, compared to the rearing of other livestock, particularly in its use of marine (and other) feed resources, are relevant.
The dietary necessity for both n-3 and n-6 fatty acids for proper development, the health of the vascular system, and the brain, has long been known. The importance of including the nutritional value (to humans) of aquaculture products in such equations was emphasized in a paper by Crawford et al. (1999). These authors, noting that the African savannah ecosystem of the large mammals and primates was associated with a dramatic decline in relative brain capacity, showed that this was associated with a decline in docosahexaenoic acid (DHA) from the food chain. The richest source of DHA is the marine food chain, while the savannah food chain offers little. In their study, Crawford et al. (1999) found that blood cholesterol, blood pressure, and lipoproteins are lower in Africans living on the shores of Lakes Turkana and Nyasa, compared to their vegetarian cousins on the savannahs, and to Europeans. Differences in blood cholesterol and blood pressure can be observed in European children living in East Africa as young as 6 years old, whose levels continue to rise, while those of the Africans remain stable. This paper provides a potent and recent example of the nutritional value of fish.
Other factors will affect the accuracy of the projections on fishmeal and fish oil usage in aquafeeds provided in this study. These include:
Chinese aquaculture and its future demand for aquafeeds. It is clear that all the projections are very much influenced by current trends in Chinese aquaculture production (as are all forecasts of expected expansion in global aquaculture production). If these trends do not continue it will have a very marked effect on the global scene.
Changing species composition as aquaculture production expands. Another important consideration is that the numerical trend analysis in this study is based upon the species that are currently being grown. China, in particular, is already exhibiting a tendency to introduce and cultivate many new carnivorous species. If this trend continues, it will undoubtedly affect all predictions about Chinese aquaculture production and its need for marine feed resources. It is not inconceivable that, during the period under consideration (30 years), China may become a major producer of non-indigenous species of marine fish, such as various flatfish, for example. Such developments would certainly affect Chinas demand for fishmeal and fish oil. There may also be significant changes in the species composition of fish reared through aquaculture in other parts of the world.
Consumer resistance. Problems related to consumer perception of farmed fish and crustaceans fed on products from the capture fisheries industry have already been discussed in earlier parts of this section of the paper. It will also be essential to ensure consumer acceptance of any proposed replacements for conventional marine ingredient resources, whether they be of animal or plant origin. Consumer choice, rather than the volume and price of marine resources, may prove the limiting factor for carnivorous fish and crustacean farming. It is possible that the original concept that a fishmeal trap might constrain certain forms of aquaculture, which was primarily an economic consideration, may be joined by other traps, such as the BSE trap, or the GMO ingredients trap. The replacement of fishmeal will probably occur less rapidly in developing countries than in developed countries. Environmental and ethical concerns, as well as economic factors, are likely to become important more rapidly in developed countries.
 However, speciality items
are derived from some of these species. In Iceland the roe of capelin is
extracted and exported, mainly to Japan, and the rest of the fish is converted
into fishmeal. Elsewhere, capelin is normally supplied whole to the fishmeal
 The conversion from raw fish to fishmeal and oil depends on species and seasons. The average for fishmeal is a yield of between 22 and 25% - the latter percentage is obtained when the contents of stick water is recovered - while the yield for fish oil fluctuates considerably from 2 to 12%. In this case a recovery of 25% has been used for fishmeal, i.e. four tonnes of fish provides one tonne of meal.
 The calculation is as follows: costs other than raw fish are calculated to be US$ 551 - US$ 367 = US$ 183/tonne of fishmeal produced. If the fishmeal price were to increase to US$ 1 100/tonne as a result of growing demand, in extreme cases the price of raw fish could increase to US$ 229/tonne (= [1 100 - 183]/4).
 Including the culture of sea trout, which accounts for about 10% of total production.
 International Fishmeal and Fishoil Organisation.
 Researching Alternatives to Fish Oil in Aquaculture.