improvement in TGC and FCR increased profit and decreased environmental impacts because improving TGC increases production and improving FCR increases production efficiency. This relationship between EV and ENV calculated from mass- based functional unit was also observed in dairy farming (Bell et al., 2011; van Middelaar et al., 2014). In chapter 5, however, when investigating the EV and ENV(fish) of TGC and FCR in sea cage system, we found that the synergy between the two values was weaker than in RAS. This was because, in sea cages, the fixed environmental impacts represented a lower proportion of the total impacts which decreased the dilution of environmental impacts over higher production. In chapter 6, we compared the genetic gain when using EV or ENV(fish). We found differences especially when the quota was on Qstock and on Qdaily_feed. These differences in genetic gain in TGC and in FCR lead to a lower economic response when using ENV(fish) and a lower decrease of eutrophication when using EV. There are differences in the response because the ENV(fish) of TGC was rather small due to the fact that fixed environmental impacts represented a small proportion of the total impacts. Therefore, we can imagine that in a system where fixed environmental impacts are larger, such as RAS, the ENV(fish) becomes also larger, which would generate the same trends in genetic gain of TGC and FCR when using EV or ENV(fish). These results suggest that it would be easier to develop breeding programs increasing profit and decreasing environmental impacts per unit of fish in farming systems with higher proportion of fixed environmental impacts, such as RAS. Additionally, in systems with more variable environmental impacts, such as sea cages, improving traits affecting production efficiency should be emphasised.
7.4. Enhancing sustainability in fish farming
Improving production efficiency
Improving production efficiency always increased profit and reduced environmental impacts, because improving the production efficiency reduces the amount of inputs (e.g. feed) used to produce one kg of fish, which reduces both the economic and environmental costs of production. In addition, improving production efficiency is also of interest for reducing the feed-food competition. The feed-food competition refers to the use of human edible ingredients for feeding animals. The improvement of production efficiency, while keeping production level and the diet constant, allows for reducing the use of human edible ingredients for feeding animals as it decreases the total use of feed per kg of animal produced. Consequently, selective breeding should focus on traits that contribute to better
158
production efficiency rather than higher production. This is especially true when it is not possible to develop different breeding programs for different production systems because any type of system would benefit economically and environmentally benefit from an improvement of production efficiency.
Production efficiency can be expressed at different levels. In this thesis, we looked at a change of production efficiency at the fish level, because the change in FCR reduced the amount of feed required by the fish to reach harvest weight. Nonetheless, selective breeding for FCR in fish farming is currently challenged by the ability to estimate individual variation in feed intake. This trait is, difficult to measure because fish are kept in groups in tanks. Consequently, genetic parameters for feed intake and FCR are lacking and selection for more efficient fish is still difficult. Different methods to estimate feed intake have been developed. One method is based on using body weight variations during feeding deprivation and re-feeding periods as predictors of residual feed intake, a proxy of feed efficiency. Using this method, Grima et al. (2008) showed, in rainbow trout, that combining weight loss during feed deprivation and compensatory growth during re- feeding period could explain about 60 % of the variation of residual feed intake. Daulé et al. (2014) performed one generation of divergent selection in sea bass based on weight loss during feed deprivation but they did not find any significant response in feed efficiency. However, the authors argued that a second generation of selection could reveal differences in feed efficiency, as in the base population, sea bass with a lower weight loss during fasting also had a lower residual feed intake. Another way to improve FCR could be by selecting on lipid deposition, which in pigs, was shown to be correlated with feed efficiency (Hermesch et al., 2000; Gilbert et al., 2007). However, there are inconsistencies in the results of experiments with. In rainbow trout, Quillet et al. (2007) showed no significant correlation between muscle fat content and feed efficiency after two generations of selection for high or low muscle fat content. Though they found a significant difference after 7 generations of selection (Quillet, pers. comm.). However, they estimated feed intake and feed efficiency in groups of fish, which cannot provide individual performances. Conversely, Quinton et al. (2007) showed for European whitefish that whole body lipid percentage displayed positive phenotypic and genetic correlations with growth rate and feed intake. This finding suggested that direct selection on growth rate together with an indirect selection against lipid content could be a way to improve feed efficiency. In this study they used X- radiography to detect feed pellets marked with dense marker in the gastro-