Link Aquaculture - Project Completion Form

Restricted – Commercial (When completed)

LINK AQUACULTURE - PROJECT COMPLETION FORM

Project Ref. No.Project Start DateProjectCompletion Date

ENV12 1/10/1999 30/9/2002

Name of LINK Project: Development of data based models for effective treatment and the environmentally safe use of veterinary methods in the control of sea lice infestation of farmed salmon.

Name of LINK Programme: AQUACULTURE

Please enter details in the boxes below:

Project Leader’s Name:: Prof. George Gettinby

Organisation: University of Strathclyde

Address: Department of Statistics and Modelling Science

Livingstone TowerAddress:

26 Richmond Street

Glasgow G1 1XH

Tel:0141 548 3805Fax: 0141 552 2079 Email::

Project Costs (£)/Effort
Government Department / Research Council / Industry / Total
Approved Spend / 143,722 / 161,400 / 305,122
Actual Spend / 143,694 / 161,400 / 305,094
Approved Staff Input* / 3 years
Actual Staff Input* / 3 years

* Staff years

Participants

List all project participants by the following categories:

Industry

Research Base

Other

Large Enterprises /

Small and Medium Sized Enterprises*

Higher Education Institutes

/ Other Research Base Partner /

Other

Marine Harvest (Scotland) Ltd. / Scottish Quality Salmon / University of Strathclyde

* Less than 500 employees and with an annual turnover of less than £30 million

Section 2 – Objectives

Audit of Marine Harvest McConnell (MHM) databases.

Identification of risk factors which lead to high levels of infestation by application of risk quantification and Hazard Analysis Critical Control Point (HACCP) techniques.

Mathematical modelling using deterministic methods such as differential equations and stage development matrix/network models for host-parasite relationships, and the use of stochastic simulation models for life-cycle analyses. (A request to emphasise the statistical modelling of the database in preference to extensive work on formal population modelling was made at the 1st Annual Review and was supported by the Project Monitoring Officer and Programme Co-ordinator – 12/2/2001.)

Validation of models using prospective data collection, sampling and data-analysis methods including the design of future minimum optimal data sets.

Assessment of different treatment modalities using multifactorial experimental design and Taguchi methods for the identification of environmental factors that minimise sea lice burdens.

Milestones

/ Target Date /

Milestones Met?

Number

Title

In full

On time

01/01 / Commence literature review / 1/10/99 / Yes / Yes
01/02 / Commence database audit / 1/10/99 / Yes / Yes
01/03 / Identification of external data / 1/1/00 / Yes / Yes
01/04 / Complete literature review / 1/4/00 / Yes / Yes
01/05 / Complete database audit / 31/9/00 / Yes / Yes
01/06 / Extend data source acquisition complete / 31/9/00 / Yes / Extended
01/07 / Dissemination of results / 31/9/00 / Yes / Yes
02/01 / Modelling methodologies assessed / 1/4/00 / Yes / Yes
02/02 / Risk assessment commences / 1/4/00 / Yes / Yes
02/03 / Prospective data collection commences / 1/7/00 / Yes / Yes
02/04 / Modelling assessment completed / 1/4/01 / Yes / Yes
02/05 / Risk assessment completed / 1/10/01 / Yes / Extended
02/06 / Data collection completed / 1/1/02 / Yes / Extended
02/07 / Dissemination of results / 31/9/01 / Yes / Yes
03/01 / Host-parasite modelling commences / 1/1/01 / Yes / Yes
03/02 / Host-parasite modelling completed / 1/7/02 / Yes* / Yes*
04/01 / Model validation commences / 1/10/01 / Yes / Yes
04/02 / Model validation completed / 1/7/02 / mostly / na
05/01-2 / System interventions initiated/assessed / 31/9/02 / Yes / Yes
05/03-4 / Delivery environment developed/assessed / 31/9/02 / No / na
05/05 / Dissemination of results / 31/9/02 / Yes / Yes

* amended target (agreed 1991)

Executive Summary of Research and Results (Include targets and objectives indicating progress made towards achieving them, or reasons for those not achieved. Also include highlights, outputs, deliverables and any unexpected benefits)

Objective 1:Audit of Marine Harvest McConnell (MHM) databases

Initially an extensive review of the sea lice literature was carried out, followed by an audit of the data collection processes adopted by the industrial partner in order to identify data that would be most appropriate. This led to the compilation of the SULLepsiS (Strathclyde University LINK Leps in Salmon) industrial health records database on treatment details and L. salmonis counts from over 88,000 fish randomly sampled during production from 40 commercial fish farms located on the West Coast of Scotland over the period 1996 to 2000. This database has been augmented with relevant environmental, production and farm site data sets to produce a single integrated resource on sea lice infestations during a period when a wide range of veterinary medicines were used to control sea lice infestations under commercial conditions. Entries to the database have been screened for consistency and completeness, with areas of uncertainty being resolved by site visits and discussion with health managers.

Objective 2:Identification of risk factors

This objective was focussed on identifying key epidemiological environmental factors which predispose sites to high levels of sea lice infestation. Over the years there has been considerable speculation as to what factors might affect the abundance of sea lice, much of which is based on limited evidence. Statistical analyses using the SULLepsiS database was used to examine the effect of yearly production stage, season and year, stock type, region, coastal exposure and level of treatment on sea lice infestation levels. Our results showed that there was tremendous variation in sea lice infestation patterns from year to year; and that stock type, geographical region and coastal exposure did not affect mean levels of abundance. In contrast treatments led to pronounced cycles of sea lice infestations with peaks and troughs at 3-week intervals, and these interventions were necesssary if the sea lice levels on fish were to be controlled. There was no evidence of water temperature affecting the mean annual abundance of sea lice infestation. As pointed out at the second annual review risk quantification using Hazard Analysis Critical Control Point (HACCP) techniques were not found to be appropriate and further risk assessment would be best pursued using classical regression methods. This work is now reported in Objectives 4 and 5.

Objective 3:Mathematical modelling

Several host-parasite population models were formulated during the course of the research programme and a population simulation model of the life cycle of L salmonis was demonstrated. Despite these activities providing proof of concept and being areas in which progress could be made, following the end of the second year a request to emphasise the productive statistical and epidemiological modelling of the database in preference to extensive work on formal population modelling was made and granted by the Project Monitoring Team. This change in emphasis has been successfully carried out. However, some unexpected progress on formal modelling has been made, using time series modelling on observations of the lesser-studied C. elongatus, in four farms on the west coast of Scotland over the period 1996 to 2000. Least Squares and Poisson regression techniques were applied to one individual site and to the aggregated data of four sites. Models were fitted and tested for goodness-of-fit using appropriate statistical methods. Findings indicated that infestation levels were highly seasonal with rapidly increasing numbers after week 22 followed by a steady decline from week 40 till the end of the year. Abundance was much lower in the second year of the production cycle than the first. Neither of the models indicated that treatment application had a significant effect on C. elongatus infestations, an observation supported by the epidemiological analyses in Objective 2 using national level data from the SULLepsiS database .

Objective 4: Validation and optimal design

From the statistical, mathematical modelling and descriptive epidemiological findings undertaken in the first three objectives an overall picture of the relationship between the two major sea lice species and their farmed environment emerged. In particular, it was evident that C. elongatus was a major species during the first year of fish production yet its epidemiology has not been extensively studied. In order to validate these observations further data were collected consisting of the abundances of the adult stage of C. elongatus in salmon populations from 33 farms on the west of Scotland between 1997 and 2000 and analysed for evidence of seasonal and annual patterns. The findings indicated that the pattern of C. elongatus was remarkably consistent from year to year, and directly opposed to that reported for L. salmonis the other major species. In particular, adult infestations rapidly increased from the start of July each year, and were more prevalent on salmon in the first year of production than the second year. Treatment was seen to have an effect on levels of infestation but it was not clear why this species should have significantly lower levels of abundance in the second year of production. There was also evidence that fallowing had no effect on abundance. Strategic management programmes for the control of sea lice on salmon farms, which are increasingly effective in controlling L. salmonis, should give greater consideration to C. elongatus.

The design of future minimum optimal data sets was investigated using an expert review panel for cause and effect analysis . Seven health managers selected on the basis of their years of experience of managing sea lice control at sites throughout Scotland took part in a workshop . Following a short presentation and discussion of the epidemiological patterns of infestation on salmon in recent years, each member was invited to draw up two lists of as many as ten factors they considered to have a substantive effect on sea lice abundance. A Taguchi approach of classifying factors into control and noise parameters was taken. One list consisted of those factors regarded by the managers as under the control of the farm management system. The other list contained those factors outwith the control of farm management. In practice, this enabled the managers to focus on both management and environmental factors. The managers were asked to rank the factors on each list in order of importance and all lists were then compared. Following discussion of the ranks the expert panel arrived at a consensus opinion on the number and the format of the factors to be included in more rigorous quantitative analyses as reported in Objective 5.

Objective 5: Multifactorial experimental design and risk factors

This is probably the most important deliverable from the project. The mean mobile abundance for three important six-month periods within the production cycle have been analysed for significant risk factors associated with increased abundance. Using multivariable statistical regression techniques over twenty management and environmental variables suspected by health managers to have an effect on controlling lice populations were investigated as potential risk factors. The findings and models developed provided a picture of mobile L. salmonis infestation patterns on Scottish farm sites collectively. The results identified level of treatment, type of treatment, cage volume, current speed, loch flushing time and sea lice levels in the preceding period to be key explanatory factors. Factors such as stocking density, site biomass, water temperature and the presence of neighbours which previously have been cited to be important correlates of sea lice risk from analysis of individual sites over time were not found to be important. The variation in mobile abundance in the first half of the second year of production could be adequately explained (adjusted R2 between 55% and 72%) by the recorded data, suggesting that there was scope for management to control L. salmonis abundance though much of the variation still remains unexplained.

Key epidemiological findings to emerge from the research programme:

Sea lice infestations are a serious threat to the viability and health of salmon farming in the UK.

L salmonis is the major species, and is prevalent during first and second year production cycles on most UK farms.

L salmonis levels are difficult to predict due to their variation from site to site, from year to year and also from season to season.

The factors stocking density, biomass, water temperature and the presence of neighbouring farms have not been found to be risk factors for L salmonis levels.

The factors level of treatment, type of treatment, cage volume, current speed, loch flushing time and previous sea lice levels have been found to be risk factors for L salmonis levels.

Stock type, geographical region and coastal exposure do not affect means levels of abundance of L salmonis.

L salmonis mobile levels rapidly recover following treatment.

L salmonis levels are controlled by veterinary methods which are administered in proportion to challenge.

C elongatus is an important species during the first year of production.

C elongatus follow a predictable time series pattern from week to week during each of the production cycles.

There is no evidence of treatment having any serious impact on C elongatus levels.

The behaviour of the two major species C elongatus and L salmonis species are distinctly different and as yet the degree of interaction and competition remains them remains unclear.

Summary of progress and achievements

The three-year work programme has been productive with a large number of key findings being communicated to the scientific literature. In some cases the findings have reaffirmed those previously reported but the substantial database established within the research project has meant that findings could be expressed with confidence. More important, a large number of beliefs and claims about the role of other factors influencing sea lice infestations on farmed salmon populations could not be substantiated.

In addition, the work has involved commercial partners and so the findings on those environmental and management factors which can have an effect on sea lice populations on farmed salmon have found their way back to the industrial partners.

The work programme was found to be most effectively undertaken using statistical modelling techniques of the substantial SULLepsiS database established during the course of the project. The establishment of the database took up a large proportion of the project time but this proved to be an excellent resource as a great deal of information emerged. The population and time series modelling aspects of the project were also productive and there is a need for these to be further pursued now that the epidemiology of the sea lice species on salmon has become clearer. A further historical database on sea lice counts infrom the late the early 1970s?? has been preserved in electronic form and the findings it contains will be the subject of future work, along with further analysis of the SULLepsiS database. The success of the project has largely been brought about by the excellent links with the staff of the industrial partners and the close cooperation with other members of the sea lice research community.

Dissemination

In addition to the publications in peer-reviewed journals listed in Patents and Publications section, a number of scientific presentations were made at:

- Aquaculture 2000, SECC Glasgow (poster presentation)

- CFRD workshop on Sea Lice Monitoring Protocols, Pitlochry 2000

- Society of Veterinary Epidemiology and Preventive Medicine conference, Amsterdam, March 2001

- Salmon Farming: Towards an integrated pest management strategy for sea lice, Aberdeen, June 2001

- SQS Technical Seminar, Perth, February 2002

- Aquaculture 2002, SECC Glasgow (multimedia and poster presentation)

- International Congress of Parasitology X, Vancouver, August 2002

The project also contributed to the establishment of the ISLM (Integrated Sea Lice Management) group, with participation and presentation at quarterly meetings, and inputs to the industry guideline Managing Resistance in Sea Lice.

In summary, the work has provided findings which can be used at national level for the strategic management of sea lice populations on farmed salmon. As yet we do not have a tactical decision support model which could provide advice at individual site level for farmers. This level of decision support will need further research and represents a goal for the future.

Executive Summary of Research and Results/continued:

Future R&D resulting from this project (Include any other non-tangible benefits and state any Teaching Companies Scheme action if appropriate)

)

In view of the findings to emerge from this study a new research and development programme has been proposed and submitted. This new programme will build upon the present findings but also be more broadly based. Key objectives of the new programme will be:

To further develop and validate the statistical linear models for the identification of farm management and environment factors that can reduce levels of sea lice infestation on farmed salmon.

To investigate the application of novel methods for the sustainable control of sea lice and in particular the role of the use of new in-feed veterinary medicines within an integrated pest management strategy.

To establish a network simulation model of local sea lice populations during salmon production cycles that can be used to support tactical on-farm decision making.

To formulate mathematical population models of the temporal relationship between the two major species of sea lice on salmon hosts, taking account of potential interaction, competition and differences in life-cycle behaviours to assess the effectiveness of national control strategies.

To review historical records of various detailed sea lice patterns on farmed salmon between 1976 and 1980 for evidence of temporal changes in sea lice epidemiology.

Industrial relevance and plans for future commercial exploitation

The findings from the current project have been widely promulgated. In addition there has been close collaboration with the industrial partners who have made use of the scientific findings in determining practice and policy for the control of sea lice infestations and the use of veterinary medicines.

It is clear that the findings contrast with that reported for other countries and so it is proposed to undertake comparative epidemiological studies of sea lice control practices commonly adopted in Europe, North America and Chile to ensure best practices are adopted within the UK.

The epidemiological and modelling approaches associated with large databases have not been widely adopted in aquaculture and so future work will promote the use of these generic analytical methods emerging from the research programme to ectoparasitic diseases of other commercially farmed fish species.