Assessing the interactions and effects of sea lice on wild salmon populations: final report of project FW0050

Summary of progress in relation to aims

1. Migration routes and distributions of emigrating salmon smolts

Tracking of salmon smolts to determine their behaviour and habitat use in coastal and inshore waters is key to understanding the potential for exposure and interaction with planktonic populations of sea lice. For this reason, work on tracking smolts was undertaken within FW0050 and further expanded in collaboration with the Atlantic Salmon Trust (through the West Coast Tracking Project) and the MD’s offshore renewable energy work. Tracking salmon at sea involves capturing smolts as they migrate down rivers and tagging them with miniature acoustic tags. The signals from these tags are then detected and decoded in receiver listening stations placed in the marine environment. For the first two years of FW0050 this work focused on smolt migration through Loch Linnhe with fish tagged on the Lochy and Awe catchments. This work determined that smolts did not have a preferred route around the Isle of Mull situated the end of the loch, travelling either north or south around the island. It also indicated that some fish could change direction during their migration moving from the Sound of Mull, into the Firth of Lorne and vice versa. This behaviour did not appear to be associated with near surface water currents.

A small scale tracking study was conducted on the river Applecross, which opens to a small embayment rather than a sea loch. This suggested that migrating fish moved using a northerly/north westerly trajectory when there were no geographical impediments to their travel. Subsequent simulation modelling of smolt movements, coupled with smolt behaviours indicated that salmon could reach the shelf current, necessary for their transport to oceanic feeding areas, by using a combination of current-following and a northerly heading (Ounsley et al. 2019).

In Loch Torridon a gridded array of acoustic receivers were employed during 2018 and 2019 to record smolt migratory behaviour through this loch system. This study indicated that some fish could move through the basins of the loch relatively quickly but some smolts also displayed milling behaviours that delayed their exit to the sea. Subsequent years (2022, 2023, 2024) of tracking in this loch were conducted to try and determine whether tagging effects and/or genetics could explain these behaviours. These studies are currently being analysed prior to being written up for publication.

1.1 West coast tracking project

The west coast tracking project, led by the Atlantic Salmon Trust, was funded for 3 years fieldwork, initially by the European Maritime and Fishery Fund and then by the MD with additional contributions by the Atlantic Salmon Trust and Salmon Scotland. The project represented a partnership between the Atlantic Salmon Trust, the MD and Fisheries Management Scotland with additional input from researchers based at the University of Glasgow. The aim was to provide data on the large-scale migration pathways and speed that salmon smolts from different west coast rivers use through sea lochs and the wider coastal environment toward the continental shelf. In the first year of study (2021), the work focused on placing receiver lines across the Minch to identify whether smolts from a number of west coast river catchments (including the Awe and Lochy) could be detected moving northwards through the Minch, or whether they followed a westerly vector, beyond the Outer Hebrides to reach the shelf current (Rodger et al. 2024). Focus then moved to mapping migration behaviours of salmon smolts around the Isle of Mull, and other complex loch systems and in the final year documenting migration speed through a variety of different loch types. This work is being analysed and written up for publication. Further information can be found here.

Atlantic salmon trout west coast tracking project

2. Variation in salmon lice levels across a range of temporal and spatial scales

In areas of salmon farming activity, levels of planktonic sea lice can become elevated with farms becoming the primary source of lice in the coastal environment. Therefore, to examine lice levels in areas with aquaculture, estimation of lice numbers on farms was collated using publicly available data. These data were initially published by Salmon Scotland (from 2011), as part of an agreement for openness through the Fish Health Working Group, and then in more detail at Aquaculture Scotland by Scottish Environment Protection Agency (SEPA) in support of the developing Sea Lice Risk Assessment Framework (SLRAF).

Prior to FW0050, MD scientists had assessed lice infestation on farms by using sea lice treatment rates, and this continued into early FW0050 (Murray 2016a). However, this indirect approach fell away, partly because non-chemical treatments were increasingly being used by the industry, such as cleaner fish and mechanical methods (so assessment based on chemotherapeutants use became less meaningful) but also because more granular data on sea lice numbers became available.

Adult female lice data had begun to be published for regional averaged counts from 2011 and so early in FW0050 it was possible to publish an initial analysis of lice loads (Murray 2016b) and, as more detailed data were published, further analysis was possible (Hall and Murray 2018). Site-level data became available from 2018, and this was used in an analysis of farmed salmon welfare indicators during Covid (Murray et al. 2021). These data indicate that there has been a decline in adult female sea lice numbers per fish on farms over time since 2015 (Rabe et al. 2024).

Sea lice data obtained from wild sea trout collected by Fisheries Management Scotland and funded by the MD was published (Sea lice on wild sea trout data). These data were analysed in conjunction with sea lice numbers on farms and confirmed that increased infection of wild fish is associated with increased adult female lice counts on farms (Ives et al. 2024).

3. Spatial and temporal variation in the efficacy of various anti-salmon lice treatments

Analysis of medicinal treatments for sea lice was carried out early in FW0050 (Murray 2016a), and indicated that there had been an increase in their usage coupled with a decrease in efficacy prior to 2011. Subsequently, medicinal treatment reduced and there was an adoption of new mechanical (e.g. thermolicers) and biological (cleaner fish) controls. This complex mix of control measures of different efficacy and form is challenging to analyse. As developers will consider a range of farmed fish health and welfare matters prior to selecting the most appropriate treatment the MD supported a project to assess cost effectiveness of alternative sea lice control methods (Boerlage et al. 2024) outwith FW0050. Work on risk factors for elevated lice levels has been undertaken and a presentation made (Ives et al. 2023), this continues to be refined, Study of the emergence of resistance to treatment (Lipschutz et. al. 2019) was supported by the MD.

The use of cleaner fish has been associated with the risk of disease emergence in salmon and this risk has been modelled by the MD (Murray 2016c, 2017). This work has been used to support advice on the reuse of cleaner fish in aquaculture. In addition, modelling demonstrated that the amount of treatment required to control sea lice can be reduced by the use of area fallowing and this was effective even when moderate larval lice exchange occurred across boundaries (Murray and Salama 2016).

Access to a range of treatments remains important consideration for sea lice management in Scotland.

3.1 Assessing sea lice dispersal in the environment

A key area of work in FW0050 has been on dispersal modelling, which is needed to assess transport of larval lice between populations (Salama et al. 2018, Murray et al. 2022, Moriarty et al. 2023). The assessment of a lice dispersal model using sentinel net pen data (temporary cages holding small numbers of fish placed in strategic locations) demonstrated the feasibility of this data collection method for model validation (Salama et al. 2018, Pert et al. 2022, Rabe et al. 2024) and was further developed under SPILLS (see below). The dispersal model was also used to assess averaged kernel transport patterns (Salama et al. 2016) which estimated averaged distances of sea lice concentration distributions from a source. General patterns of pathogen dispersal at different scales of time and space were reviewed (Murray and Salama 2017) and the benefits of area management modelling assessed (Murray and Salama 2016). Similarly, the dispersal modelling was used to provide advice on area management of aquaculture (Murray and Gubbins 2016), contributing to the development of EU marine spatial management policy (Gimpel et al. 2018) and to the FAO/World Bank as a case study on aquaculture zoning.

Sea lice models have been collaboratively assessed in a range of projects with different partners in which the MD has played a role. These collaborations have allowed pooling of resources and the exchange of ideas that enhance scientific endeavour and in turn strengthen the outputs of FW0050. Modelling, and gap analysis, taking into account views of a range of different organisations and sectors, ensured that concerns of different stakeholders were included in the models and that there is widespread buy-in to the approach. This collaboration included specific mechanisms for alternative models developed by different groups of modellers to be compared. The following are a list of projects conducted during FW0050 that involved the collaboration of scientists working at the MD and which supported the objective of sea lice dispersal model development.

LiceTrack

LiceTrack was sponsored by the North Atlantic Salmon Conservation Organisation (NASCO) – using European Union (EU) funds – to establish sea lice modelling for lice management in Killary Harbour, Ireland. In addition to the MD, the project involved Norwegian Institute for Marine Research (IMR), Inland Fisheries Ireland and the National University of Ireland Galway. This project aligned with FW0050 objectives and the MD provided advice both on modelling and on the use of sentinel cages to collect data. An output of particular use for FW0050 was the development of an agreed model framework through international authorship that describes the requirements a valid sea lice model must incorporate (Murray et al. 2022).

SPILLS

The project Salmon Parasites in Linnhe Lorne and Shuna (SPILLS), was funded by Scottish Aquaculture Innovation Centre (SAIC) with in-kind contributions from Mowi and other salmon aquaculture industry stakeholders. In this project the MD, Scottish Association for Marine Science (SAMS) and Mowi tested their models against an historical data set, collected by the MD, to share lessons, optimise models in use, and identify model sensitivities to uncertainties in model structures and available data. SPILLS project final report

SLUG

Sea Lice Uncertainty Group (SLUG) is a collaboration between the MD, SAMS, IMR (Norway) and Firum (Faroe Islands) that has sought to describe the causes of uncertainties in model outputs to advise on sea lice management. Therefore, identifying gaps that can be filled and if not, the implications best described, so contributing to the understanding of “knowledge strength” by stakeholders (Murray et al. 2025). In addition, the project created an Evidence Map listing the gaps and the evidence needed to fill gaps. The aim is to provide the best advice in an imperfect world. Marine Directorate Information page for SLUG

SAVED

Sustainable Aquaculture Validated for Ectoparasite Dispersal models (SAVED) was a project led by Strathclyde University with SAIC funding to create a model validation framework (SAIC project homepage). The project was established at a SLUG meeting by the MD and Strathclyde University and includes a range of stakeholders from Scotland, Norway and the Faroe Islands to establish this framework using Scottish data that can be shared by modelers seeking to validate their own sea lice models. In addition, modelling of infection processes was undertaken (Waite et al. 2024). An Ensemble approach whereby multiple models are used to generate a consensus was initiated in this project, and continues to be developed.

AquaDEEP

A Defra Seafood Innovation Fund (SIF) funded project in which a sea lice dispersal model was developed by BMT for a decision support prototype, the MD provided advice and shared lessons from the model’s development.

SLIPD

The MD provides scientific input to a SEPA funded project called Standardized Lice Infection Pressure detection (SLIPD). This project funds the IMR, Norway to develop approaches for sentinel cage use to maximise data collection of estimates of lice to salmon infestation pressure.

Specific to FW0050, the collaboration with external modellers via the LiceTrack project allowed model components to be codified (Murray et al. 2022). Formal development of model structure and a gap analysis were conducted (Moriarty et al. 2024, Murphy et al. 2024). Specific model improvements were examined, with the detailed infection processes in the immediate vicinity of hosts driven by swimming of copepodids and smolts being included (Murray and Moriarty 2021, Moriarty et al. 2023). Infection processes have also been modelled in SAVED (Waites et al. 2024). The inclusion of diurnal vertical migration into models showed its importance in driving local accumulations of infectious copepods stages (Garnier et al. 2024). Dispersal modelling was further advanced through developments made using international collaborations in the SPILLS and SAVED projects. As an outcome, models were used to inform aspects of the SLRAF in discussion with SEPA.

3.2 Oceanographic models applied to sea lice dispersal

Dispersal models for planktonic sea lice are dependent on oceanographic models of currents, and the MD has developed the Scottish Shelf Model (SSM), which has become a key tool for various applications. The use of coupled hydrodynamic – particle dispersal models has been supported by the establishment and publishing of good practice guidelines for dispersal models (Garnier et al. 2025). The SSM model allows for multiple specific scenarios that simulate a particular year, together with scenarios that describe typical climate averages for assessing average dispersal for long-term planning. Highly resolved sub-domains of the model can be used to cover the coastal areas in which aquaculture occurs and this has continued in collaboration with SEPA and the National Oceanography Centre to inform the SLRAF. A specific application of the SSM during FW0050 has been to the modelling of sea lice dispersal over large distances between different regions (Rabe et al. 2020). This initiative allowed farm connectivity to be assessed, which is useful for developing area co-ordinated management strategies.

4. Infestation pressure experienced by salmon in relation to numbers of salmon lice in environment

Data on the presence, absence and numbers of sea lice in plankton samples obtained from known localities is critical for the validation of models to understand sea lice dispersal and impact. The MD collected data on sea lice in Loch Linne in 2011, 2012 and 2013 (available at Biological sampling data in Loch Linnhe), prior to the initiation of FW0050. These data have subsequently been used in validation for modelling developed in FW0050 (Salama et al. 2018) and in the associated SPILLS project.

Further sentinel cage work was carried out in Loch Linnhe during FW0050, including in the development of prototype towed sentinel cages, allowing lice to be sampled along trajectories (Pert et al. 2023). The MD supported the collection of new data by others related to sentinel cages through working with SEPA and their funded partners (see SLIPD project).

Data on planktonic lice was collected in Loch Shuna by Scottish Association of Marine Science (SAMS) during the SPILLS project Sea lice data Shuna sound. In addition, a subsample of routine planktonic trawl data collected by the MD from two marine locations (one west coast and one east coast) was compared to provide an indication of lice levels in an area with aquaculture and an area without (Harte et al. 2017).

During FW0050, the MD reviewed sea lice surveillance methods (Pert et al. 2022, Rabe et al. 2024) and this work continues to support an advisory role for the MD in the standardisation of sentinel net pen design as part of the SLIPD project.

Modelling of the relationship between concentration of lice in the environment was also undertaken from first principles of interaction (Murray and Moriarty 2021, Moriarty et al. 2023) and from observed distributions on sentinel cages (Waite et al. 2024). However, the infection process is identified as an area for further work (Murphy et al. 2024).

5. Impact of salmon lice on individual salmon.

A meta-analysis of existing studies on the sea lice impacts on salmon was conducted. This work identified a threshold of 0.08 lice g^-1 fish for welfare and behavioural impacts and 0.24 lice g^-1 for a 50% probability of mortality (Ives et al. 2023). These thresholds are consistent with those identified elsewhere (e.g. Taranger et al. 2015) and are used to support parameter values in SLRAF.

6. Effects of anti- salmon lice treatments on population losses of salmon released into the wild around the coast.

In 2015, the Scottish Aquaculture Research Forum (SARF) funded a complementary project to FW0050 to determine the feasibility of quantifying the impact of sea lice on salmon returning to natal rivers. The approach was based on an approach used in Norway and Ireland of releasing smolts in groups that either have or have not been subject to sea lice treatments to identify effects of lice on survival by comparison of the groups. The SARF project was to inform the potential for adoption of a network of treatment sites on the east and west coasts of Scotland, therefore providing direct assessments of sea lice impact associated with aquaculture.

The plan for Scotland was cognisant of conservation issues regarding the use of hatchery reared fish and so wild smolts were caught and treated. The rivers chosen were the Conon draining into the Cromarty Firth on the east coast, and the rivers Loy and Lundy draining into the Lochy at the head of Loch Linnhe on the west coast. The treatment chosen was Substance Ex- an unlicensed treatment for which lice have no known resistance. Generally low return rates, and difficulties in operating temporary traps on the west coast resulted in a determination that the approach was not suitable for widespread adoption (Morris et al. 2018). However, a subsequent, small-scale study, was conducted using a fish lift fitted with a passive integrated tag (PIT) reader in the hydroelectric dam at Loch Awe. This study eschewed the issue of operating temporary traps in spate rivers and suggested a possible way by which this work could be conducted at a west coast site in the future if required (Morris et al. 2019).

6.1 Effects of sea lice infestation pressure on a sea trout population

To examine the impact of sea lice on sea trout, the MD collected data at a permanent fish trap set up on the Shieldaig river. The study was initiated in 1999, and continued through FW0050 until 2020. To supplement low numbers of wild fish the river was stocked with hatchery-reared fish, marked by the removal of the adipose fin. The study recorded the size of sea trout leaving the river, and the numbers and size of the fish returning. Sea lice infestation pressure was estimated by examining lice levels on sea trout electro-fished in the short stretch of river below the trap toward the sea. From the analysis of the data it was concluded that the marine growth of sea trout was adversely affected by sea lice. As the infestation pressure increased, so the growth rate of the returning fish decreased. Wild fish had a greater return rate than the hatchery reared fish at all levels of infestation pressure. There was strong evidence of an effect of infestation pressure on the return rates of hatchery fish, but this effect was not significant at the 0.05 significance level for the wild fish (Morris et al. 2026).