Reducing Flavobacterium psychrophilum transmission risk via hatchery-rearing practices: An individual-based modeling evaluation

Abstract

Flavobacterium psychrophilum can be a significant mortality source in hatcheries where trout and salmon (Family Salmonidae) are propagated and raised. Preventing F. psychrophilum outbreaks and controlling its spread in hatcheries is challenged by limited effectiveness of current therapeutics and lack of efficacious commercial vaccines. Consequently, one of the few available options to limit outbreaks and spread within a facility may be modifications to hatchery-rearing practices. We developed an individual-based model to evaluate how rearing practices could limit F. psychrophilum-induced losses of Oncorhynchus mykiss in flow-through, partial-reuse raceway systems. Evaluated practices included frequency of removal of dead individuals, the probability that hatchery personnel found fish that died from infection, and water flow rate through raceways. Practices were assessed under different assumptions about initial number of infected individuals and virulence levels associated with infection. Increasing the frequency of removal of dead individuals led to the largest reductions in cumulative mortality due to disease as well as reductions in peak daily infection rates and time until peak daily infections occurred when compared to changes in flow rates. Under high initial infection, high virulence, and low flow conditions, increasing the frequency of removal from once to two or three times per day decreased cumulative mortality by approximately 150 thousand fish. When initial infection prevalence was high, disease outbreaks were acute with peak daily infection rates occurring 39 to 53 days post-placement in outdoor raceways. Conversely, when initial infection prevalence was low, infection course was more chronic. Increasing the frequency of dead fish removal to hourly and the probability of hatchery personnel finding dead fish in raceways when checks were performed resulted in modest gains in survival, although these results were likely conditional on model assumptions. If the probability of hatchery personnel finding dead fish was lower than what was assumed in our evaluations, larger gains from more frequently checking for dead fish may have resulted. We believe there is considerable potential in using simulation models to explore the benefits of different rearing practices in cultured fish. To that end, we have made the simulation model used in this study available for those interested in adapting or expanding it to address other research questions.