Science of salmon stocking: report

4. The development of a stocking programme

Careful specification and application of a stocking programme is required to prevent doing more harm than good (Cowx, 1994b; Aprahamian et al., 2003; Naish et al., 2007; NASCO, 2007; Araki and Schmid, 2010; Young, 2013). The successful implementation of such a programme requires the development of a strategic approach that identifies the problem, defines the objectives, orientates the implementation to meet the goals, and effectively monitors the outcomes. The various steps for consideration in setting up such a programme are set out below and outlined in Fig. 2 (Cowx, 1994b; Cowx et al., 2012).

Figure 2 Factors for consideration when setting up a stocking programme (modified after Cowx et al., 2012; Anon, 2013).

Non-biological considerations

• Legal constraints
• Policy considerations
• Socioeconomic factors
• Business case

Non-biological decisions

• In line with international/National/local legislation?
• Supported by policy?
• Sectoral/commercial/public interest?
• Positive business case outcome?

In no to any of these factors then reject stocking

Ecological considerations

• Environmental impact
• Competitive interactions
• Genetic considerations
• Disease & parasite risks

Ecological decisions

• Any negative environmental impacts?
• Any negative impacts on wild fish?
• Any negative impacts on wild fish of other species?
• Any unmanageable genetic risks?
• Any unmanageable disease/parasite risks?

In no to any of these factors then reject stocking

Hatchery considerations

• Hatchery performance
• Hatchery running
• Hatchery environmental impact

Hatchery decisions

• Can hatchery fulfil requirements?
• Are there resources to staff hatchery?
• Is hatchery footprint, water use, waste disposal and biosecurity manageable?

In no to any of these factors then reject stocking

Implementation considerations

• Financial planning
• Construction
• Hatchery staff expertise
• Operational planning

Implementation decisions

• Is finance available for duration of project?
• Can construction impacts be managed?
• Are personnel available with required skills?
• Is there a full hatchery operational plan in place covering all aspects of running and stocking?

In no to any of these factors then reject stocking

Develop operational plan

• If stocking to take place then operational plan to be developed to include:
• Broodstock collection
• Crossing scheme
• Stocking plan - numbers, ages, timing, location
• Monitoring

4.1 Non-biological considerations

A stocking programme, as with any management intervention, must be performed legally. Planned activities must thus be compliant with any relevant legislation at the international, national and local levels. Such legislation may cover all stocking of fish in a country (e.g. the complete prohibition on stocking in Wales: Natural Resources Wales, 2014) or specific activities at specific times in specific places. It may not even be directly based on stocking of salmon, but rather on the impact of such activities on other protected species and/or habitats. It is vital, therefore, that a thorough evaluation be made of the relevant legislation.

Similarly, although not defined in primary legislation, there may be policies applied by regulatory bodies and/or management organisations which are relevant to a planned stocking programme. Again, these may cover specific activities/times/places but also the aims of any proposed programme. For example, there may be a ban on stocking for certain purposes (e.g. enhancement) while other types may be allowable or even encouraged (e.g. conservation). Such policies and associated guidance must again be carefully considered before any programme is implemented.

In many historic situations, socioeconomic drivers have been of particular importance when setting up a stocking programme. Societal value and economic income from salmon fishing in many areas has been high and, with recent declines in the available resource, the drivers for management intervention are also high. Such pressure must, however, be balanced by the best scientific advice available to try to achieve agreement between the drivers and the science, as well as consensus between those on differing sides of the debate. Achieving such a consensus is not, however, straightforward as was well illustrated by the discussions surrounding the recent stocking policy development in Wales which resulted in entrenched and antagonistic positions being developed by different groups of stakeholders (outlined in Harrison et al., 2019b).

Finally, a stocking programme must have a comprehensive business case. The funding, planning, aims, operation and monitoring must be well defined at the outset. Such a business case does not have to be financially viable in the same way as a traditional one and, indeed, a stocking programme may be a financial loss-making enterprise even if the full economic benefits are calculated, yet still be a conservation success. It does, however, have to be based on a sound financial plan for operation throughout the lifetime of the planned intervention, whether this is through regulatory body funding, direct cash injection by interested parties, fundraising, and/or other avenues. Full costs should thus be defined at the outset and yearly breakdown of operational costs and available funds outlined at the planning stage.

4.2 Ecological considerations

As with any anthropogenic environmental intervention, careful consideration must be made as to the ecological risks of such actions over and above that of the focus stocks (Holmlund and Hammer, 2004). In order to assess likely impacts and as part of the plan development for the stocking project, initial ecosystem-wide analysis of the site of stocking would be required to examine the potential effects of the action on both wild conspecifics and other species taking account of competition and predator/prey interactions. Through the evaluation of this analysis, site-specific issues could be addressed and mitigation solutions, as well as, monitoring strategies brought into the plan.

4.3 Hatchery considerations

Establishment of a stocking programme necessitates the development of a hatchery operation. Examination of the full implementation and running of such an establishment is beyond the scope of this review. However, it is imperative during such development that proper consideration is given to the aims of the programme and the ability of the hatchery to meet these aims. There is little point in developing a hatchery before a thorough stocking plan has been developed, which will include a definition of the required numbers of fish, the capacity to keep separated different genetic groups of fish, and the ability of the hatchery to meet these requirements. It is also imperative that the staffing and other resource implications are defined and the environmental impact of running the hatchery is carefully considered. This will include disruption during building and maintenance of the physical structure, and also, importantly, impacts from the discharge of the facility (Michael, 2003; Tello et al., 2010). Fish and food waste will require monitoring and biosecurity of both fish and potential diseases and parasites established (Lillehaug et al., 2015).

4.4 Implementation considerations

It is perhaps obvious that a hatchery development, as with any other business, requires a detailed financial plan (Cowx et al., 2012). Such a plan should cover both the construction and running of the facility, with full costings of infrastructure and staffing for the duration of the project. A hatchery represents a significant investment and may mean substantial capital input and ongoing running costs. It is thus vital that a realistic commercial plan is put in place to cover the financial aspects of the operation. The running of such a facility requires particular expertise and, again, there should be a plan in place to either obtain or develop the required skills.

4.4.1 Operational plan

The running of a hatchery supplementation programme requires a detailed operational plan covering all aspects of the intervention. The plan should cover the various practical aspects of the operation (outlined in Fig. 3) and be reviewed regularly to ensure the operation is following the planned trajectory and/or if the plan requires alteration in the light of new information. In order to ensure best practice, it would perhaps be helpful, at both the planning and review stages, to take independent advice and oversight from outside experts.

Figure 3 Factors for consideration when planning a hatchery operation.

Broodstock

• Collection point/s
• Collection timing
• Fish ages and sexes
• Sizes
• Numbers
• Potential contamination

Hatchery fish production

• Crossing scheme
• Maximise genetic diversity
• Hatchery generations
• Replenishment of hatchery lines
• Production of sterile fish

Hatchery Conditions

• Water requirements
• Diseases/parasites
• Densities
• Habitat enrichment

Fish Releases

• Numbers
• Ages
• Sizes
• Locations
• Timing

Monitoring

• Species of interest
• Other species
• Wider ecosystem
• Socioeconomics

4.4.2 Broodstock

The hatchery will be utilising wild-caught broodstock and careful consideration is thus required to determine the impact of removal of these fish from the wild population. It is often the case that supplementation programmes are undertaken when wild stocks are low or falling. In such cases, care must be taken to avoid significant additional negative impacts on the stocks remaining (McElhany et al., 2000). In the absence of an identified stressor but with the wide body of evidence showing the reduced fitness of hatchery-produced fish compared to wild (Jonsson et al., 2019), natural spawning should be prioritized where possible. While, in some cases where specific stressors have been identified (e.g. wild habitat loss, ecological pollution, parasite impacts in the wild) (Gausen, 1993; Hesthagen and Larsen, 2003; O'Reilly and Doyle, 2007; ICES, 2018; Soininen et al., 2019), the most favourable option may be to remove broodstock fish to the hatchery, in many other situations (e.g. where conditions in the wild are still favourable for freshwater production and the system is not at carrying capacity), it will be best to maximise wild production by allowing natural spawning.

A further factor of importance when evaluating the impact of broodstock removal from wild stocks is the structural relationships between populations and sub-populations which can change over time as stressors such as marine survival vary both within and between years (ICES, 2019). This may influence how broodstock selection can be organised. i.e. in poor periods of survival sub-population diversity is likely to fall as the smaller sub-populations are impacted by increased influences of straying from larger adjacent populations (Consuegra et al., 2005). Such impacts may result in a loss of significant population differentiation and a homogenisation of sub-populations. In turn this may mean that is becomes possible to increase the extent of areas available for collection of broodfish without deleterious effect to the existing new population structure. However, below a certain critical threshold, populations are very sensitive to any changes. Removal of members of a small sub-population may take them over the brink making it unviable and leading to possible extinction of that sub population. Care must thus be taken before such an approach is attempted and carful analysis of the sub-population structure undertaken.

Following consideration of the ability of the wild stock to tolerate removal of broodstock fish for the hatchery, the next step is to ensure that collection is carried out in such a way as to maximise the fitness of the offspring produced. Determining where to collect the adult fish is sometimes simple; for example, if a system has a dam preventing upstream movement then stocks can only be taken as close as possible below the dam. This does not mean however, that the fish will actually represent fish from the population/s of focus, due to a degree of straying that occurs during adult return spawning migrations (Malcolm et al., 2010; Keefer and Caudill, 2014). If the barrier is towards the head of the river and/or on the main stem, fish from many tributaries and/or rivers may be captured, with subsequent hatchery production resulting in homogenisation of regionally differentiated, locally adapted populations (Williamson and May, 2005; Östergren et al., 2021). Interbreeding between fish from different locally adapted populations can lead to negative effects on survival and fitness of the offspring, a process defined as outbreeding depression (Fraser et al., 2010; Houde et al., 2011).

In cases where there is no single barrier to migration, the decision as to where to collect fish is not a simple one, due to the many locally adapted populations that may be present in a system (Taylor, 1991; Garcia de Leaniz et al., 2007) members of which may be migrating through different points in the system at different times (Stewart et al., 2006). Fish taken from the mainstem of a river may capture much of the genetic variation in a system. However, the production of crosses with such fish will result in mixing of stocks with, again, the associated risk of loss of local adaptation through outbreeding depression (Manhard et al., 2018). At the other extreme, fish taken from a single/small tributary may miss much of the available wild genetic variation, and subsequent crossing may result in loss of fitness through inbreeding depression (Frankham, 2005). The degree to which either inbreeding and/or outbreeding impacts fitness may be highly unpredictable, even at small genetic distances. As such, there is a need to evaluate the relative risks of inbreeding and outbreeding on a case-by-case basis (Houde et al., 2011).

A second factor to take into consideration is when to take broodstock from a system. Fish return to their natal spawning grounds throughout the year and these adaptive genetic differences associated with run timing are population-specific (Vähä et al., 2011; Cauwelier et al., 2018a). Therefore, unless collection is on the actual spawning grounds during spawning season, there is a danger of missing stock components if all fish are taken at a single time point. In order to capture the full component of the stock, it might be the case that multiple collections take place throughout the year. This does, however, raise the potential problem of shortage of the required number of fish, if numbers are still required late in the season but no fish happen to appear at this time. As such, a careful analysis of the migratory timing of the population of interest should be undertaken and a plan developed to attempt to maximise variation in time of return in order to prevent unintentional selection away from natural patterns (Ford et al., 2006).

A further complexity to the collection of broodstock is that populations are comprised of fish of different ages and sexes (Palstra et al., 2009). In order to maximise genetic diversity, consideration should be made as to the representation of these different groups in the broodstock. The age a fish returns to breed has a significant genetic component (Ayllon et al., 2015; Barson et al., 2015; Aykanat et al., 2019). Fish of different sea ages should thus be collected in proportions similar to the wild stocks and sex ratios matched to known spawning activities (e.g. Taggart et al., 2001; Jones and Hutchings, 2002). Further, as precocious male parr are known to fertilise a significant proportion of eggs (Saura et al., 2008), they should also be collected and used as broodstock. This will maximise the number of breeders to produce the next generation and result in a large effective population size (Garcia-Vazquez et al., 2001).

Knowledge of the local biology of the wild fish when collecting broodstock is vital for capturing and maximising natural variation, thereby ensuring that the stocked offspring represent the wild population as much as possible. However, anglers like to catch big fish and so it is tempting to select the largest fish when choosing broodstock. However, this act alone will produce an unnatural directional selective pressure on the stocks, as it is well known that size is important in mate choice in salmonids (Fleming, 1996; Auld et al., 2019) and has evolved to an optimum in a particular population (Jonsson et al., 1991b; Roni and Quinn, 1995). Thus, again, in order to maintain genetic diversity, a representative selection of fish size at age should be used.

Following determination of where, when and what to collect as broodstock, sufficient numbers should be collected to match the aims of the programme with regards to production but also to minimise the reduction in genetic variation that results when a small subset of a larger population is used as broodstock (founder effects) (Frankham, 2010; Witzenberger and Hochkirch, 2011). This may seem an obvious step, but to avoid situations where either not enough fish are available to meet conservation goals or excess fish are removed from the wild, planning is required to match requirements to objectives. Careful consideration should be given to the range of fish to be collected and the numbers of eggs likely to be produced from fish of a given size using length/fecundity relationships derived from the stocks of interest. Only then can evidence-based broodstock numbers be scientifically justified.

In many areas, a particularly important consideration for hatchery managers when utilising wild-caught broodstock is the risk of contamination of hatchery lines through the use of broodstock that have either themselves escaped from aquaculture facilities or are hybrid offspring of wild fish and farm escapees. It is imperative that broodstock are screened to ensure the fish are not contaminated by aquaculture stocks in areas where there may be impacts from escaped farm fish. There are genetic tools available that can distinguish between farm, wild and hybrid fish (Karlsson et al., 2014; Gilbey et al., 2018; Wringe et al., 2018). Such techniques can be performed relatively rapidly and, so, could be employed wherever broodstock are retained, even if only for short periods. In other cases, where fish are captured and immediately stripped, post-crossing evaluation could be performed and groups of eggs produced from contaminated parents destroyed.

4.4.3 Hatchery fish production

Approaches to achieve the goal of preserving genetic diversity in the hatchery depend on the goal of the supplementation programme. There are two components to preserving genetic diversity: (1) maximising effective population size, and (2) using non-random mating to increase the diversity of genotypes above that expected from random mating (Fisch et al., 2015). The crossing scheme employed in any particular situation should thus be carefully considered. Significant research has been undertaken on this subject (reviewed in Fisch et al., 2015) and considerations include whether to use random or non-random mating, whether to mate single pairs only or use a factorial design, whether to monitor relatedness to avoid sibling-crosses and/or minimise inbreeding, whether to allow free mate choice, and whether to equalise family sizes. The final crossing design decided on will thus have to match available broodstock numbers with conservation requirements and resources available for pedigree monitoring and/or fish rearing. As with the other aspects of setting up a supplementation programme, a detailed crossing scheme plan should be produced before operations commence, following careful review of the available scientific guidance.

Consideration should also be made with regards to the duration that broodstock will be kept in captivity, as well as whether a number of offspring from subsequent generations will be used as broodstock. This is important, as it is well known that, when a stock is retained in a hatchery situation, it is subjected to domestication selection and associated loss of fitness in the wild (Fleming and Einum, 1997; Lynch and O'Hely, 2001; Frankham, 2008; Fraser, 2008) and that, though such fitness reductions can occur within a single generation (Christie et al., 2012a; Milot et al., 2013; Christie et al., 2016), the more generations a stock has been under this selection, the lower the fitness in the wild becomes (Berejikian and Ford, 2004; Araki et al., 2008; Christie et al., 2014; Minegishi et al., 2019). Taking such observations into consideration, it is clear that, if possible, stocking should take place using F1 offspring of wild-caught broodstock. This will reduce any negative fitness impacts of captive rearing. However, in some cases, for example if there is limited wild broodstock, using subsequent hatchery generations may be justified. Again, plans should be developed on a case-by-case basis, depending on the available resources and best scientific advice for the specific situation and supplementation aims.

In certain stocking situations, where, for example, the aims are to enhance stocks for recreational and/or commercial exploitation, consideration should also be made to the possibility of producing and stocking with sterile fish. Sterile (triploid) fish can be produced by heat shocking (Crozier and Moffett, 1989) or pressure treating (Kozfkay et al., 2005) eggs and is now used widely for stocking various species of trout (e.g. Scheerer et al., 1987; Chatterji et al., 2007; Scott et al., 2014). New gene-editing techniques are also becoming available to produce sterile fish (e.g. Dankel, 2018) that may also become of use in the future (especially in hatcheries producing fish for aquaculture purposes). The stocking of such fish will remove the potential direct genetic impact of hatchery fish breeding with wild stock and associated significant negative impacts. However, indirect genetic effects, through mechanisms, such as competitive interactions (Santostefano et al., 2017), mean careful consideration is still required.

4.4.4 Hatchery conditions

Running a hatchery brings the usual practical considerations around water use, waste management, disease and parasite control and biosecurity of fish. The aims of the hatchery are to produce fish that have maximum fitness when released into the wild. As such, consideration should be made into the conditions under which the fish will be raised and whether measures can be undertaken to maximise such fitness. There is much evidence that raising fish in an 'enriched' environment leads to enhanced fitness when released into the wild (Flagg and Nash, 1999; Hyvärinen and Rodewald, 2013), as an enriched environment leads to impacts on numerous physiological processes (Crank et al., 2019 and references therein). Enrichment possibilities include but are not restricted to: 1) enhancing habitat complexity by providing matrix substrates and darkened environments; 2) promoting development of body camouflage coloration by creating more natural environments, such as overhead cover and in-stream structures and substrates; 3) conditioning fish to bottom orientation by positioning food delivery low in tanks; 4) altering water flow velocities to exercise fish and so enhance predator avoidance; 5) improving foraging by supplementing diets with natural live foods; and 6) adjusting rearing densities to more natural spatial distributions (Flagg and Nash, 1999 and references therein).

4.4.5 Fish releases

Important considerations for any stocking programme are when, where, how many and at what age to release hatchery stocks. Again, such decisions will be based on an interaction of conservation aims, river system characteristics and practicalities. An important decision for successful stocking is determining age-specific carrying capacity, natural production and setting a corresponding stocking rate (Solomon, 1985). Matching numbers of stocked fish with available resource is critical in order to both promote maximum production from the system and fitness of fish but also to prevent negative effects on wild stocks through impacts on resource availability (Aprahamian et al., 2003 and references therein). Age-specific juvenile carrying capacity can be estimated (e.g. Malcolm et al., 2019) and location specific stocking densities matched to this after taking into account natural production.

The age at which fish should be stocked is, again, another balance between survival and resources. It is obviously much cheaper to stock fish at an early stage, perhaps even as ova, rather than to go through the expense of rearing them to juvenile or smolt stages. The younger the fish are at stocking, the better their adaptation to the wild is likely to be (less domestication selection) and so the smaller the difference in individual expected post-smolt survival compared to wild fish will be (Salminen et al., 2007). However, this intrinsic benefit is offset by the fact that rearing fish in a hatchery to later ages avoids the extremely high natural mortality that occurs during early life stages in the wild (Ware, 1975; Gee et al., 1978; Good et al., 2001). Survival of stocked fish of different ages and life history stages varies greatly (see Table 9 in Aprahamian et al., 2003) and so, again, analysis should be performed and related to the specific goals and resources available to the supplementation exercise.

Although the size of stocked fish will be associated with age, it is also important that fish are stocked at appropriate age-specific sizes. Wild juvenile salmon size-at-age has evolved to maximise fitness in local conditions (Swansburg et al., 2002) and size is important in defining many developmental life history traits, such as precocious maturation (Baglinière and Maisse, 1985; Thomaz et al., 1997) and smolt age (Metcalfe et al., 1989; Heggenes and Metcalfe, 1991; Pearlstein et al., 2007). Stocking of often well-fed large size-at-age fish thus has the potential to disrupt natural competitive interactions in the wild. However, there may also be logical reasons to stock with fish at larger size-at-age than wild fish, depending, again, on the aims of the supplementation programme. If, for example, stocking is with sterile fish in a put-and-take fishery such as with trout, then anglers may want to catch larger individuals. Also, as early salmon post-smolt mortality is high (Thorstad et al., 2012; Chaput et al., 2018) and strongly influenced by size (Salminen et al., 1994; Saloniemi et al., 2004; Jokikokko et al., 2006; Gregory et al., 2019), it may be possible to boost hatchery fish survival by the release of larger better conditioned smolts (Saloniemi et al., 2004).

A further consideration for any supplementation programme is where to stock the fish. In some cases, this may be determined by particular constraints (e.g. dams or other barriers), whereas, in others, careful planning should be undertaken so as not to negatively impact wild recipient populations. It is vital that the carrying capacity of the system is taken into consideration when planning where to stock (Kelly-Quinn and Bracken, 1989). Small numbers of fish stocked uniformly across systems or in areas of known low local densities will thus enhance the likelihood of positive outcomes. However, even if such careful considerations are taken, the strong natural juvenile density-dependent mortality may mean positive impacts may not be realised (e.g. Glover et al., 2018).

The timing of releases is also of importance in ensuring maximum production from the programme. Stocking in the spring has been found to be 4-12 times more efficient than in the winter (Cresswell, 1981; Aass, 1993). Ideally, fish should be released when temperatures and flow are relatively low and productivity is high, but not during spawning period (Cowx et al., 2012). Stocking in spring and/or early summer is thus preferable to allow fish to acclimatise before overwintering. Together with considerations as to whether to stock at a single point (spot planting) or distributed more widely (scatter planting), enhanced success may be obtained if fish are introduced using 'trickle planting' (i.e. releasing batches of fish over an extended time period) (Cowx, 1994a). Trickle planting can reduce competitive interactions by reducing over-dispersion of released fish and, as such, scatter and trickle planting should be preferred over spot planting (Berg and Jorgensen, 1991; Fjellheim et al., 1993; Cowx et al., 2012). As with other aspect of a stocking programme, however, the resources to enable stocking over extended periods must be balanced by the likely enhancements in success rates through such actions.

4.4.6 Monitoring

Once a stocking proposal has been accepted and, as part of the project plan development, all projects should have the methodology in place to enable adequate monitoring of progress and, ultimately, success or failure of the intervention (Cowx et al., 2012). While the objectives of a stocking programme should be clear and set out in the project plan, it is often the case that the lack of suitable monitoring programmes means the efficacy of the programme and the ability to detect impacts or attribute improvements directly to the stocking is lacking (Cowx, 1994b; Waples et al., 2007; Bacon et al., 2015; Glover et al., 2018). Monitoring is important, not only to inform managers of the efficiency of the programme underway but also to inform the development of future programmes of a similar nature.

In order to effectively monitor a stocking programme a baseline is required to which any changes may be compared. Ideally this should cover a number of years and so be able to capture the variations in the various metrics under investigation. That said , certain circumstances, such as imminent extinction, may require a more pragmatic approach to collect the best data achievable in the time available. Cowx et al. (2012) outlined a series of factors to be included in a stocking monitoring programme, which should run over an appropriate time-scale and include technical, ecological, genetic and social considerations. They suggested these should cover:

• changes in production trends of stocked/resident fish species
• changes in the genetic integrity of stocked/resident fish species
• changes in growth performance of stocked/resident fish species
• changes in species and catch composition
• impacts of latent disease and parasites
• impacts on the habitats (e.g. loss of aquatic vegetation, changes in the composition of aquatic vegetation, increases in dissolved solids and turbidity) of recipient ecosystems
• impacts on the trophic dynamics of recipient ecosystems (e.g. changes in the quality and quantity of plankton communities, increases in single age groups of particular fish species, changes in the quality and quantity of benthic organisms)
• changes in the socioeconomic conditions of people related to the fisheries

Monitoring programmes should thus seek to determine not only how well ongoing supplementation meets the original aims of the project with regards to the species of interest but also look into the impacts on both the wider ecosystem and the socioeconomics of the people and communities involved. Such a monitoring programme will enable decisions to be made at strategic points in the programme as to whether to continue as planned, revise plans and/or stop the inputs altogether, if either negative impacts are detected or realised outcomes are not meeting projections. As with all aspects of the programme, a science-based monitoring programme should be included as part of the project proposal.