Effects of Feed on Water Quality in Recirculating Aquaculture Systems

Fish production in recirculating aquaculture systems (RAS) has been expanding rapidly over the last decade. The business model of RAS farms and increasing environmental regulations require RAS production to be intensified with a high fish density and a low water exchange rate, which lead to the accumulation of fish waste excretion. To ensure fast fish growth and good fish health, water quality should always be kept optimum. In pursuit of this, we need to comprehensively describe different factors that influence water quality. Fish feed is the primary source of nutrient input into RAS. Manipulating diets (e.g., feed loading and dietary composition) has been proven to affect the quantity and composition of fish waste excretion. However, the knowledge of dietary effects on RAS water quality is still scarce.

This Ph.D. study aimed to examine the effects of feed on both physicochemical water quality and microbiology in RAS. To this end, four separate trials were conducted with special focuses on dietary elemental composition and feed dust, using pilot-scale RAS with rainbow trout (Oncorhynchus mykiss) as model systems.

The objectives of the first two trials were to understand the relationship between dietary phosphorus (P) levels and orthophosphate-P (PO₄-P) accumulation in RAS and the associated effects on physicochemical water quality and microbiology. In trial 1, three diets with increasing total P (0.89%, 1.20%, and 1.49%) were administered to rainbow trout in triplicated pilot scale RAS for 5 weeks. The results demonstrate that PO₄-P accumulation in RAS correlates positively to digestible dietary P when the available dietary P exceeds the minimum requirement of the fish. In trial 2, two diets and external tri-sodium phosphate were administered in quadruplicate to achieve three levels of PO₄-P concentrations in RAS. The results showed that microbial activity in system water correlated positively to PO₄-P concentration while microbial community compositions in system water and biofilter biofilm (including nitrifier groups) were insensitive to different PO₄-P concentrations in the range tested.

The objective of trial 3 was to test the potential effects of different dietary C:N (carbon:nitrogen) ratios on physicochemical water quality parameters and on RAS microbiology including nitrifier communities. In this trial, four experimental diets with increasing C:N ratios (8.5, 9.3, 10.0, and 12.2 by weight/weight) were tested in triplicated RAS for 4 weeks. The dietary C:N ratios spanned the range found in commercial rainbow trout growth-out diets. The best water quality was found in systems administered the diet with the highest C:N ratio, in terms of significantly lower nitrite-N (NO₂-N) and nitrate-N (NO₃-N) concentrations, lower dissolved organic matter concentrations, reduced microbial activity, and high water clarity. The results, on the other hand, showed that manipulating dietary C:N ratio did not lead to changes in C:N ratio in system water. The overall microbial communities in system water and biofilter biofilm differed from each other, whereas none of them responded to changes in water C:N ratios induced by the diets, presumably due to microbes being highly C-limited in all groups. However, the relative abundance of ammonia-oxidizing bacteria was positively correlated to dietary protein content.

The objective of trial 4 was to test the effects of feed dust on physicochemical and microbial water quality in RAS. Four levels of feed dust (0%, 2%, 4%, and 6% of the feed) were tested in triplicated pilot RAS with rainbow trout for 4 weeks. The results showed that particulate organic matter concentrations and microbial activity were positively correlated to feed dust proportions. The concentrations of total ammonia nitrogen (TAN) and NO₂-N were only moderately affected by feed dust, while the concentration of dissolved organic matter was unaffected.

The first two trials highlight the potential benefit of using low P diets in RAS, i.e., reducing PO₄-P accumulation and P discharge and controlling microbial activity without disrupting the stable state of microbial community. The third trial shows that manipulating dietary C:N ratio can affect physicochemical water quality in RAS while microbial communities seem to be robust against changes in water quality induced by the dietary changes. The fourth trial substantiates the potential deteriorative effects of feed dust on particulate organic matter and microbial activity in RAS. Overall, the results in these four trials substantiate that diets can affect water quality and microbiology in RAS.