Fish farm health evaluation: interpretation of site mortality records

Abstract

In aquaculture worldwide, diseases are a significant constraint to economic expansion. The Scottish salmonid industry has experienced many cycles of development, with episodes of little or negative profitability caused by excess of production, and times of crisis due to different disease problems. In Scotland, the early implementation of regulation largely contributed to the control of infectious disease outbreaks. The recent Chilean outbreak of infectious salmon anaemia (ISA) illustrated the threats and the impacts of disease in the aquaculture industry and the importance of implementing good regulation and husbandry practices to reduce the impact of the spread of infectious disease. Databases of site production data have an important role to play in the investigation and understanding of diseases. They store valuable data collected during the time of production, which are essential for the identification of potential health and production problems during the production cycle of farmed fish. Mortality records are one of the most important sources of information on a farm, especially if it includes the cause of death as deformities, predators and diseases. Any deviation from the expected levels of mortality may indicate production problems, infectious diseases, or inadequate welfare. The investigation of increased rates of mortality must include examining farm records, determining the influence of death rate on production and the potential risk factors of diseases in a farm. This project demonstrated the importance of mortality records for setting industry standards of “expected” mortality losses and for investigating the value of recorded mortalities as a tool for aiding in surveillance and control of infectious diseases. It also aimed to determine the utility of reported mortality in supporting and assisting management-strategy decisions at the farm and industry level. In this project, we developed a baseline benchmark curve for expected mortality losses for Atlantic salmon in seawater. This novel approach constitutes a first attempt to establish a baseline curve for normal mortality, which allows detection of potential production problems based on deviations of mortality from the baseline curve of normal mortality. The results of this study also indicated that mortality levels may vary across production cycles, which can again be identified by using the baseline. We found that site was the factor with the highest contribution to variance in mortality. This site-to-site variation in mortality may have resulted from epidemics and environmental incidents, or other local event/effects. Temperature, and/or geographical area were also characteristics that contribute to variation in mortality. The regulator, Marine Scotland Science, with the backing and support of the salmonid industry has suggested potential mortality thresholds as an indicator of presence of infectious diseases, which could be used as alerts for inspection by the official authority. In this study, high mortality rates on fish farms were investigated as an indicator of the presence of infectious disease. The analysis was performed using several analytical approaches: receiver operating characteristic (ROC) curve analysis, measures of sensitivity and specificity, and bootstrap methods. The study was performed by splitting the production cycle into small fish with mean weight below 750 g and large fish with mean weight over 750 g. In the small fish, the results did not suggest reported mortality as a strong indicator of the presence of infectious disease, which may be caused by the lack of records of infectious disease at this stage of the production cycle. In the larger fish, high mortality rates were found to be a strong potential indicator of the presence of infectious diseases, including the suggested mortality threshold. In a survey, the role of traditional diagnosis in the prevention and control of disease outbreaks was assessed. For that, key informant interviews were performed with open questions to the health or farm manager of several trout and Atlantic salmon farms and we also used the diagnostic reports of the Veterinary Diagnostic Services (VDS) from Stirling University to triangulate the data. We showed that disease diagnoses are of great importance for disease identification and control of actual diseases. Farmer’s experience was also indicated as essential in the identification of the first signs of disease, which was principally through the daily monitoring of fish. This study suggested that disease diagnosis starts at the farm level with the daily monitoring of fish and the records of different parameters by the farmer, including mortality. Those records were showed to be vital to identify problems within the production. This thesis illustrated a novel approach to investigate and interpret recorded mortality at the farm level. The results presented in this thesis indicated reported mortality as a vital on-farm tool for identification of diseases and production problems. This thesis suggested priority areas where further investigation is required.