Towards robust predictions of salmon lice infestation pressure in Scottish Lochs: Biophysical model parameter sensitivity testing

Future growth within the Scottish salmon aquaculture will be dependent on its environmental sustainability, which means that, among other considerations, aquaculture operators must ensure that sea lice originating from farms do not cause unsustainable mortality to wild salmon. Management of salmon lice must be informed by a set of high-quality (calibrated) and well-documented salmon lice dispersal models. Here, we have documented the sensitivity analyses undertaken to assess the variance and fit of a set of models used by industry, government and academics for salmon lice dispersal modelling, each using different particle tracking models with the same hydrodynamical forcing in this case. Loch Linnhe, a fjordic system in the west coast of Scotland, has been used as a test area, due to available historical empirical data for model validation. The analyses carried out increased our understanding of how the model outputs which estimate salmon lice density in space and time are related to, and influenced by, the various input parameters tested. Lice swimming behaviour, specifically small scale diel vertical migration was identified as an important variable in simulating the distribution of salmon lice in the environment. The source type (used to calculate number of lice larval released from each site) influences how well the model correlates to the field data. Using the Code of Good Practice intervention threshold (0.5 adult female lice per fish) increased larval output by 10.7% compared to simulations based on observational lice data. This may be useful to managers/regulators in understanding the likely hotspots for lice in the water column, should lice loads on all farms reach these intervention levels. However, inference on the actual lice loads using a “variable” source based on lice counts provided a better fit to the field data. There is a requirement for comprehensive empirical data collection in all areas where aquaculture occurs to assess model suitability within each system.