Bycatch and discards
One source of major quantities of fish raw materials is found among what is already fished, but for various reasons is thrown back into the sea. Today’s fisheries are largely based on selective fishing where certain species are fished. In addition to the desired species, large amounts of fish are caught as bycatch. Some of the bycatch is landed and recorded, while the rest is dumped into the sea. Alverson et al. (1994) has estimated the global discarding of fish at 27 million tonnes. This means that millions of tonnes of protein are dumped annually into the ocean.

In addition to being an enormous waste of resources this practice also leads to an underestimation of world fishing pressure. For stocks that according to official landing statistics are already taxed to the maximum, bycatch with subsequent discarding will be the factor that, overall, will push stocks into an unsustainable condition.

The reasons for discarding are plainly economic. In some cases discarding will also be a result of directly illegal activities during fishing. Clucas (1997) lists several reasons for discarding from the world fishing fleet:

• Wrong fish, wrong size, wrong sex or injured fish.
• The fish cannot be stored with the rest of the catch.
• The fish are inedible or poisonous.
• The fish do not keep well.
• Lack of space on board.
• "High grading" (Low-value fish are dumped to make room for fish of higher value.)
• The quota has been reached (leads to high grading where the small fish are dumped to fill up the quota with larger fish of higher value).
• Catch of prohibited species in a prohibited area, in a period with closed fishing grounds or with prohibited gear.

According to the FAO (1997) the amount of discarded fish was reduced from the mid-1980s until the mid-1990s. The organisation is now operating with an estimate of around 20 million tonnes of discards annually. The reduction is due to a) a decline in the fishery, b) a fishing moratorium for periods/areas, c) the development of more selective fishing equipment, d) greater utilisation of bycatch for consumption and for feed for aquaculture and livestock farming, e) the introduction of a ban on discarding in certain countries and f) greater focus and willingness by governments and interest groups to reduce the amount of discarded fish.

In Norway, the authorities have adopted a zero discard policy. It is illegal for commercial fishermen to throw back any of the catch to the sea. This is an incentive to fish more selectively by avoiding fishing in certain periods and areas where high bycatches can be expected. The prohibition is also a driving force behind the development of equipment that reduces bycatches (Hall et al. 2000). The EU countries have a law that is nearly the exact opposite of Norway’s. They have introduced a prohibition against keeping, or landing, fish where a TAC (Total Allowable Catch) or quota has been reached (Alverson et al. 1996). In many cases this means that the EU requires the fishing vessels to dump fish. Most of the fish thrown back in the sea are already dead. The fish that are still living have little chance of surviving (Hall et al. 2000).

Transgenic plants (GMO)
The possibility of modifying the genes of oil-rich plants to produce a vegetable oil with a fatty acid profile that covers the needs of salmon has been aired. In addition, a substantial percentage of today’s production of, for instance, soya and maize is based on transgenic plants. The policy of the fish farming industry in Norway is to avoid transgenic plants in fish feed. The main reason is consumer scepticism towards such products. In this chapter we will show why the fish farming industry should continue its restrictive attitude, and perhaps expand the reason from market-related concerns to a real concern for considerable environmental problems.

Modern transgenic technology involves inserting genes from one organism into another. Opposition to and attention to transgenic plants has been massive, particularly in Norway and other parts of Europe, but not to the same degree in the United States. The arguments against have been that the technology goes against nature and that the consequences can be dire (Thakur, 2001). Bellona does not share the view that the technology is wrong per se, we believe the arguments should be based on the possible impact of the cultivation of transgenic plants, and the technology should be met with an extremely cautious precautionary approach.

Bellona is not generally against gene manipulation for research purposes or in controlled production of chemical compounds in secure laboratories. For example, the production of insulin by the use of transgenic bacteria or yeast cells is unproblematic. The precautionary principle must, however, apply within all the areas of use in this technology, and it is important that scientists in fields that use transgenic organisms are critical in view of the future use of research results. New technology should not be adopted before the risks to human health and environmental safety are properly evaluated. In our opinion no satisfactory environmental impact assessments have been established for releasing genetically modified (GM) organisms in the environment. This is largely due to scientists knowing too little at this time about the objective of the technology, namely DNA. It is important that scientists and other decision-makers acknowledge that knowledge about genetic material is still at a "kindergarten level". It is crucial that science has good knowledge of how genes work before GM organisms are released. This is a prerequisite for any realistic determination of long-term ecological effects. Based on the knowledge that is available today, all releases of transgenic organisms into the natural environment are unacceptable.

What are the consequences?
The problem of releasing transgenic plants into the environment is by and large associated with three factors: resistance to antibiotics, pesticides and herbicides, and Bt genes. The latter two are properties transferred to many transgenic plants because the plants will then tolerate pesticides and herbicides or will produce Bt toxin in the plant tissue, which kills certain types of insects. (Bt prototoxin is produced naturally by a group of bacteria called Bacillus thuringensis, and is converted to Bt toxin if it comes in contact with a digestive enzyme found in some groups of insects. The resistance to antibiotics is only a marker gene that with current technology can be cut out again. Antibiotic resistance genes are currently a large problem in most of the transgenic plants grown today, but can most likely be eliminated in the future.

Several problems are associated with herbicide resistance and Bt toxin genes. The Bt toxin in plant tissue that is ploughed under could have a great impact on the soil fauna and also kills species that naturally combat pests or species that have no direct effect on agriculture but otherwise have functions in the ecosystem. In contrast to traditional spraying, transgenic plants that express Bt genes constantly subject insects to Bt toxin. This will increase the probability that the pests will develop resistance. If transgenic plants with herbicide resistance cross-pollinate with wild-growing relatives, herbicide resistance could make it impossible to spray for weeds. We know that such cross-pollination will take place in several of the plants that are used, or could be used in aquaculture feed, such as maize or rape. This has been one of the opponents’ main argument, although for a long time there was no concrete evidence. In recent years, however, it has been proven that transgenic elements have spread to cultivated plants in other fields and relatives growing in the wild (Quist & Chapela, 2001).The other main problem with herbicide and insect-resistant plants is that they lead to increased and indiscriminate use of pesticides and herbicides. Because of the fact that the plants in question are resistant to herbicides and/or pesticides, an attitude has evolved that it does not hurt to spray the crop an extra time or two.

Reproduction and heredity among bacteria is very different from what we find in higher organisms. Among other things, they can absorb DNA from their surroundings and exchange small DNA fragments with each other. In this way DNA from rotting plant parts can be transferred to soil bacteria (Bertolla & Simonet, 1999) or from food in the digestive system and to intestinal bacteria (Martin-Orue et al., 2002). It was recently discovered that DNA from transgenic plants has been transferred via pollen to bacteria and yeast in the intestines of bee larvae (Kaatz et al., 2002 -in the process of being published).This shows that the artificially introduced genes are spreading in the environment, not only by normal cross-pollination with closely related species, but that they actually invade the genomes of completely non-related species.

A number of concerns are also related to some of the vectors (particularly viruses) used to transfer foreign DNA in the stem cells of the new transgenic organism. Such vectors will incorporate other genes in addition to those whose expression is desired, and could cause the transgenic element to form a new virus particle that can infect new hosts or change the expression degree of other genes.

Gene modification can also cause changes in the biochemical processes of the transgenic organisms. These can have an expression that given today’s knowledge is impossible to predict. Damaging concentration of toxins, mutagens (substances that cause potentially damaging genetic changes) and carcinogens (substances that stimulate the development of cancer) can be formed.

So far it looks as though there is limited risk associated with eating transgenic organisms. Studies, however, have shown that the digestive system can be affected by food from transgenic plants (Ewen & Pusztai, 1999). Ewen and Pusztai (1999) showed a change in the morphology of the intestines of rats fed a type of transgenic potato. When these results became known, the transgenic potato was withdrawn and further research was prohibited. It has also been demonstrated that DNA from GM crops have been transferred to intestinal bacteria in humans (Gilbert et al., 2002).The people taking part in the trial were given milkshakes and hamburgers containing GM soya with a herbicide resistance gene. This gene was subsequently found in the intestinal bacteria of the trial participants. The trial was carried out on people who have had a colostomy. It is surprising that DNA from the transgenic soya variant can survive the passage through the stomach and stay more or less intact in the small intestine.

It is also important to note that gene pollution cannot be cleaned up afterwards. The effects may thus be irreversible. Based on currently available knowledge Bellona regards transgenic plants as an unacceptable source of feed for the fish farming industry.

Fossil fish feed
Protein for feed can be produced from natural gas. Norferm at Tjeldbergodden in Norway has developed a product called BioProtein, which is now being used on an experimental scale in Norwegian salmon feed. Norferm describes the process as follows: "The production of BioProtein takes place when the microorganism Methylococcus capsulatus and a few other auxiliary organisms grow and divide continuously with a regular supply of methane gas, oxygen, ammonia, various nutrient salts and minerals. How quickly the cells grow and divide depends on the amount of nutrients added per unit of time" (www.norferm.no).The composition of the product Basic BioProtein is given in table 10.

Trials have been conducted to replace part of the fishmeal in feed with BioProtein. Various percentages of BioProtein were tested, and a diet consisting of 20 per cent BioProtein turned out to give the best result. The trials also showed that BioProtein contributed to a higher growth rate and more efficient utilisation of feed in salmon (EWOS, 2001).

Metabolising of nutrients
When nutrients are metabolised in the cells, CO2 is formed, which all animals exhale. In an environmental policy context this "emission" is not considered to be an emission of greenhouse gases because the carbon is usually part of the natural cycle. Through photosynthesis, plants take up just as much CO2 as they subsequently give off during metabolisation in the cells of animals. When part of the carbon metabolised in the cells comes from fossil sources, as is the case with BioProtein, this involves a net addition of CO2 to the system. Consequently, it must be evaluated whether this CO2 emission should be included in the accounts for emissions of greenhouse gases into the atmosphere. The CO2 emissions from the respiration of salmon do not, however, go directly into the atmosphere. Some of the carbon is bound in the flesh of the fish and subsequently released when the fish is eaten. The carbon exhaled by the fish can to a varying degree be taken up in the photosynthesis of microorganisms, algae and other plant material under water, and is bound in this cycle over time, so that a potentially extremely long-term delay in the emission occurs. There are divided opinions among scientists about how longterm such binding of CO2 is. High density of fish as found in a fish farm, will regardless lead to a higher concentration of CO2 locally. This creates higher CO2 pressure against the water surface so that CO2 is emitted into the air. The division between CO2 that is bound under water and CO2 that is emitted into the air will depend on local environmental factors such as the degree of water exchange and the amount of organisms taking up CO2. Calculation of this division at each fish farm will require extensive modelling.

Consumption of natural gas
Consumption of natural gas in the Norferm process is 2.3 standard cubic metres (Sm3) per kilogram BioProtein (Huslid, 2003). One standard cubic metre of gas (CH4) yields 2.27 kg CO2 when burned. If we assume a feed factor of 1.3, which means that 1.3 kg of feed yields 1 kg of growth in the fish, and presuppose that fish feed contains 10 per cent BioProtein, the consumption of BioProtein will be 0.13 kg per kilogram fish. The production of 1 kg BioProtein consumes 2.3 Sm3 of gas. The consumption of gas per kilogram of fish is thus: (1)0.13 kg/kg of fish x 2.3 Sm3/kg = 0.3 Sm3/kg of fish.

Combustion of 0.3 Sm3 gas yields a CO2 emission of:
(2)0.3 Sm3/kg of fish x 2.27kg CO2/ Sm3 = 0.7 kg CO2/kg of fish

To put this figure into perspective we can compare the emissions of cars. A VW Golf with a 1.6 litre petrol engine has an emission of 166 gram/km (www.volkswagen.no). Net addition of fossil CO2 to the natural cycle from the production of 1 kg of salmon is thus equivalent to the emission from a car driven 4 km.

If all Norwegian salmon receive 10 per cent BioProtein in their diet, the annual addition of fossil CO2 will be:
(3)500,000,000 kg of salmon x 0.7 kg CO2/kg of salmon = 350,000,000 kg CO2, or 350 thousand tonnes of CO2.

Similarly, 20 per cent BioProtein in the diet would yield an emission of 700,000 tonnes CO2. By comparison, total CO2 emissions from road traffic in Norway in 2001 totalled approximately 9 million tonnes (SSB, 2002). 350,000 tonnes CO2 is equivalent to 9-10 per cent of what we must reduce in relation to current emissions to meet the Kyoto obligations.

If the use of BioProtein becomes widespread, a climate policy assessment should be undertaken as to what extent a new source of fossil carbon should be permitted while simultaneously expending major resources on eliminating other sources.

Harvesting zooplankton
The research programme Calanus at the Norwegian University of Science and Technology (NTNU) is aimed at identifying the opportunities for harvesting zooplankton in Norway. The objective of the programme is to map the sustainable harvesting potential, develop efficient harvesting techniques and industrial processes for processing, and evaluate the nutritional properties of raw materials with respect to fish feed. The project must be regarded as basic research. There is extremely limited knowledge about this area at this time (Calanus – project description).

What is known, however, is that the quantity of zooplankton in the ocean is enormous. In the Norwegian Sea, the quantity of zooplankton varies between 5 and 11g dry weight per m2 (Ellertsen et al., 1999), while for the Barents Sea it varies between 8 and 13 g dry weight per m2 (Hassel, 1999).The production of different types of herbivore zooplankton, e.g. the calanoid copepod Calanus finmarchicus, is so large that if only 10 per cent of the production is harvested, it would be equivalent to the entire ocean’s biomass production in the first level of carnivorous species on the food chain including herring, capelin and carnivore zooplankton. This biomass is particularly interesting because it is highly similar to the salmon’s natural diet. In addition to small fish, salmon eat lots of copepods, e.g. Calanus finmarchicus. It is therefore assumed that Calanus finmarchicus is a well-suited source of protein for farmed salmon. At the same time this biomass is food for economically important fish, which places strict requirements on management, both in relation to ecosystem and society. The biggest practical challenge is to find cost-effective catch methods. If you can imagine a trawler with a type of pelagic net so fine-meshed that it takes copepods measuring only 1-2 millimetres, energy consumption would be dramatically high. In addition, a fine-meshed active implement has the capacity to take a bycatch consisting of virtually everything in its path. A form of passive filtering would perhaps be more realistic. Another possible approach is to produce protein by cultivating phytoplankton as a food source for copepods or other similar types of animals.

Algae as fish feed
Various types of microalgae produce fatty acids suited to the nutritional needs of the farmed fish. At the Agricultural University of Norway, a research project funded by the Research Council of Norway has been initiated to develop production methods for microalgae as a fat-rich source of feed for fish. In order for microalgae to be commercially interesting as fish feed, quick-growing algae with a high content of the desired polyunsaturated fatty acids must be selected. The production process itself must be developed so that the product can be competitive in price. Today microalgae are produced for health food purposes at a production cost of NOK 200 per kg of dry material. These costs have to drop to NOK 50 per kg for fish feed production to be profitable (Hjukse, 2003).The fatty acid composition and the potential for use in aquaculture have been studied in Duerr et al. (1998), Renaud et al. (1998) and Browna et al. (1997).