Sea lice management measures for farmed salmon production: research

6. Discussion

This section discusses results of the quantitative cost-effectiveness and life cycle analyses in the context of qualitative findings from the participatory workshop on the wider impacts of sea lice control. 

Results of the quantitative analysis indicate that sea lice management by in-feed and long term usage of skirts to prevent sea lice from entering the pens have the highest relative cost-effectiveness. Findings from the workshop analysis indicate that, according to stakeholders’ perceptions, skirts’ relatively lower impacts on environment and fish welfare are translated into positive impacts at the retail side of the supply chain and positive consumers’ perceptions. The cleaner fish, fresh water, physical removal measures and the licensed veterinary medicines were among the second most cost-effective measures, and this is supported by their mixed and at times contradictory environmental, health and welfare impacts. The use of hydrogen peroxide (both well boat and tarpaulin) represented the least cost-effective measures among single use measures and, based on the opinions of stakeholders involved in the participatory workshop, these were also regarded as less positive methods by the public in view of their fish welfare and environmental aspects, and human health implications. The results confirm the behaviour of Scottish producers who have significantly reduced the use of this measure and currently employ it very infrequently. 

Results indicate the ranking of single use sea lice management measures based on their relative cost-effectiveness show that for a measure to have the same efficacy effect (e.g. the same reduction in number of sea lice per fish), its costs may be higher than the costs of other measures. More specifically, when comparing incidental measures by the cost incurred to achieving the same level of efficacy, the use of hydrogen peroxide (well boat and tarpaulin) costs more than all other measures, while use of other licensed veterinary medicines, physical removal and fresh water measures cost less than these but more than the in-feed measure, which is the most cost-effective incidental measure. Among continuous measures, skirts were found to be the most cost-effective and ranked better than cleaner fish whose purchasing and maintenance costs were higher. As presented in the qualitative part of this study, i.e. findings from the participatory workshop, cost-effectiveness of prevention and mitigation of sea lice is not the only measure of importance as, for example, skirts are perceived to reduce oxygen flow and may have a detrimental effect on fish with compromised respiratory functions, and therefore their effectiveness concerning general fish health and welfare can be low.   

To account for the wide range of expert opinion based efficacy rates for each sea lice management measure, we run sensitivity analysis at the extreme values to identify any corresponding variation in the ranking and magnitude of measures’ cost-effectiveness. No major changes in ranking occurred under the maximum efficacy values, with the exception of physical removal (thermolicer) measure, which became less cost-effective. Under assumed minimum efficacy level, the ranking changed considerably. Skirts and in-feed measures remained as the most cost-effective measures under assumed minimum values for efficacy scores, and hydrogen peroxide remained the least cost-effective among the single use measures. Physical removal measures and use of other licensed veterinary medicines became significantly more cost-effective, while fresh water measures became significantly less cost-effective. In addition to changes in ranking, cost-effectiveness exhibits the expected changes in magnitude under both the minimum and maximum efficacy assumptions. Sensitivity analysis results indicated the variability in cost-effectiveness related to changes in efficacy levels, namely that the higher the efficacy of a measure, the higher the cost-effectiveness. 

To account for uncertainty owing to combined data sources, additional sensitivity analyses were carried out to assess the impacts of varying values for the costs of interventions on the model outcomes, such as feed conversion ratio. Changes in the costs of interventions to similar extent for all management measures did not change the overall cost-effectiveness rankings of the management measures although the absolute values of cost-effectiveness were changed. However, changing costs of single use measures showed some changes in cost effectiveness rankings. For example, a significant increase in the price of cleaner fish led to decreased cost-effectiveness of the measure per unit of effectiveness to the extent of cleaner fish becoming slightly less cost-effective than the hydrolicer measure.

Results indicate that sea lice management measures using tarpaulins were more cost-effective than measures using well boats under the whole range of efficacy values used in this study. This relates to the higher costs associated with operating well boats compared to tarpaulins. We assumed that efficacy was unaffected by the method of performing the bath measures. A study exploring differences in efficacy tarpaulins and well boats for applying licenced veterinary medicines found that efficacy by tarpaulins were 2.2 times higher than efficacy by well boat, with results constrained by limitations such as sample size and bias from abundances and sea lice stages before applying a sea lice management measure (Whyte et al., 2016). The scarcity of evidence available on the matter supports the need for investigating differences in efficacy between sea lice management measures in Scotland.

In addition to ranking single use measures based on their cost-effectiveness, we investigated a series of combination measures. These are based on realistic sequences of measures implemented throughout the production cycle to achieve acceptable efficacy by the end of the production cycle i.e. combination measures are chosen to bring down and maintain sea lice count within acceptable (regulation compliant) levels where they do not impact fish health, welfare and productivity. Cost-effectiveness of combination measures are not comparable to that of single measures as these combine different measures in different sequences, with many of them repeated and as such, aspects of cost additivity apply. When comparing the three combination measures, results indicate a small difference between the most cost-effective and least cost-effective combination ranging from £1.23 per fish per unit of effectiveness to £1.67 per fish per unit of effectiveness. Depending on farm circumstances, the difference may be even smaller, as well as the magnitude of total costs, for the cases where cost additivity can be adjusted to account for cost synergies across measures. Sensitivity analysis employed to test results for combination measures indicate high sensitivity to fish mortality. 

For modelling purposes i.e. to create a quantitatively driven ranking, the study examined individual methods, however there is typically no individual method that achieves the desired level (no/very low numbers) of sea lice across a production cycle. Hence farm managers can and do use the wide range of methods available to them, which may include those lower ranked overall, as necessary to build the optimal sea lice management strategy.

Fish welfare was taken into account in both, the qualitative workshop analysis and the quantitative models, the latter through sea lice management measure related mortalities, which is, arguably, a proxy welfare indicator. Results of the workshop showed that perceived importance of both salmon and cleaner fish welfare is high, equally for industry and consumers. Where cleaner fish are used as a sea lice control method, the welfare and health of both salmon and cleaner fish are affected by any other sea lice management measures applied. The differences between these fish species, not only in size but also many other biological parameters, imply that measures optimised for salmon may affect cleaner fish differently. Results of the workshop indicate that cleaner fish are perceived as cost-effective and their welfare is key to positive consumer perceptions. 

Participatory analysis identified potential incentives for further improving control of sea lice on farm, many of these already taken into account in the Scottish salmon sector. These include: better balancing of science-based evidence and precautionary principle based policies (health, environment, welfare); public sector driven positive incentives such as subsidised access to technology; research on consumers’ willingness to pay for sustainably farmed salmon; media campaigns and education to the public on implications of disease control in aquaculture; market based incentives (price differentiation through labelling re sustainable disease control); market based incentives (traceability); development of health monitoring/preventative technologies; development of delicing technologies from product/flesh at packing; research on salmon welfare linked to sea lice control; research on cleaner fish welfare linked to sea lice control; media campaigns to maintain/improve industry image to the public; private driven stick type incentives such as higher environmental/welfare standards required under processor/retail contracts; research on efficacy of disease control; and improved collective action to sea lice control along the supply chain.

It should be noted that caution should be taken when comparing cost-effectiveness of single use measures based on the limitations previously mentioned regarding data available for this study, data uncertainty linked to combined sources, as well as the consequent simplified modelling assumptions. Data sources are a combination of primary and secondary data, and expert opinion, moreover the geographical distribution of data sources from e.g. Norway and Canada were translated to the Scottish situation as closely as possible, but might not fully represent the Scottish salmon industry situation, more specifically as regards regional differentiation. With additional as well as more robust primary data, further ways to improve the analysis include the methodological integration of economic, biological and epidemiological modelling. The findings from the participatory workshop indicate the complexity of sea lice control not only on farm but beyond farmgate, as supply chain, regulatory and environmental effects, and the need to address it as a holistic challenge. 

Despite data limitations, ours is an analysis that adds useful insights to the current literature and creates a basis for further research. As demonstrated by the limited literature on the topic, this type of research will always be constrained by access to data for reasons detailed in this report. Thus, while much information has been collected on the variables included in the analyses and the robustness of results tested using sensitivity analysis, there may be other factors for which neither data nor robust proxies could be identified under this study, and this should be taken into consideration when reviewing these findings as a basis for future research.