Environmental transmission and connectivity of the ectoparasite Lepeophtheirus salmonis in marine ecosystems with intensive salmon aquaculture

© 2018 Dr Francisca SamsingLarge-scale epidemics are affecting ecologically and economically important marine species across the globe. For the most widely produced marine fish, salmon, parasitic sea lice are the most significant problem. Here I investigated the environmental transmission of farm-originated sea lice Lepeophtheirus salmonis in an area of intensive salmon aquaculture, and used empirical and modelling approaches to dissect the processes involved in parasite dispersal and connectivity. First, I empirically examined the effects of a wide range of biologically relevant water temperatures (3 – 20°C) on L. salmonis larval development and demonstrated that temperatures had a negative relationship with development times. My findings showed that larvae developed at a faster rate than was previously implemented in dispersal models, and changing this parameter had a significant impact on simulated dispersal patterns. In my next chapter, I used empirical data to validate the horizontal and vertical distribution of lice predicted by the dispersal model and tested different model parameterizations (changing larval development time and vertical behaviour) to see which best represented empirical data. The best model fit occurred with a development time to the infective stage that was faster than previously implemented in dispersal models, and the presence of temperature-driven vertical behaviour of lice early planktonic stages. This model also predicted that hindering fish from swimming close to the surface could reduce lice infection pressure by around 60%. In subsequent chapters, I used the validated dispersal model to construct networks of lice dispersal. In the first simulation, lice were released from all salmon farms in the southwest coast of Norway (~ 560 farms) for two seasons (spring and winter) from 2009 to 2014 to quantify spatial and temporal variations in connectivity and test the stability of dispersal networks. Lice dispersal patterns and network metrics varied greatly between seasons, but differences were consistent amongst years. Winter networks presented more connections between farms, and links were on average two times more distant than in spring (winter median connection ~ 37 km; spring median connection ~ 18 km). In the second simulation, lice particles were released from all farms in Norway (~ 1000 farms) for a whole simulation year (2016). Outputs from this simulation we used to test: 1) the best method to reduce overall connectivity; and 2) where firebreaks would be best located to disconnect the entire network. Sequential removal of key-positioned farms was more effective in fragmenting the network than removing farms with a high number of connections. I also demonstrated that the establishment of two firebreaks (at ~ 61° N and 67° N) could be achieved with the removal of 13 and 21 farms (~ 1.3 and 2.1% of all farms), fragmenting the entire network into three largely unconnected groups of farms. My work showed that dispersal models coupled to network analysis could aid future spatial planning of the salmon industry and identify strategies that keep lice infection pressure at a minimum. More broadly, the framework I used in my thesis is widely applicable to other diseases-causing agents within open-cage aquaculture in marine systems, and should be a key component in the study of marine epidemiology