Scottish scallop stocks: results of 2023 stock assessments

4. General discussion

4.1. Regional summaries

Substantial scallop fisheries have existed around the coast of Scotland for many years. While landings in the Shetland area have been systemically increasing over time (and landings at Orkney show a general increase), other areas show significant periodic fluctuations. Historically there have been occasional rapid increases/decreases in landings from some areas (East Coast, North East and North West), some of which are associated with fishery closures due to ASP/PSP toxins, but others appear to be due to the appearance of strong year classes (increases in recruitment) and an increase in stock size. In recent years (since around 2015), there have been major declines in the landings from some of the historically most important scallop areas (East Coast, North West, West of Kintyre) and a decline in total landings from the Scottish assessment areas during this period. Anecdotal information from the Scottish fishing industry suggests that this reduction is due to some of the large nomadic dredgers moving between fishing grounds and spending a significant amount of time fishing elsewhere around the coast of the UK in recent years.

The SAM stock assessments show that following periods of lower recruitment during the late 1990s, the stocks around Shetland and East Coast of Scotland experienced higher recruitment during the 2000s and 2010s resulting in increased SSB. In recent years, recruitment at the East Coast has been relatively low, with a decrease in SSB and catches. In contrast, the stock assessment for Shetland shows a rapid increase in recruitment and SSB in recent years. In the North East, after strong recruitment and high SSB in the 1990s, recruitment and SSB have been fluctuating around an average level since then. The recent reduction in catches in the East Coast area are reflected in a decrease in fishing mortality while fishing mortality in the North East is estimated to be fluctuating without trend in recent years. The fishing mortality in Shetland shows a recent decrease despite the increase in catches due to the rapid increase in SSB.

In the West of Kintyre, after low recruitment until the early 1990s, good recruitment in the 2000s and 2010s has resulted in increases in SSB. In the North West, after a period of good recruitment in the 1990s to 2000s and high SSB, recruitment and SSB have decreased in the recent period In both areas, catches have declined substantially which has resulted in a decline in fishing mortality to near the lowest in the time series.

Historical stock trends estimated by the SAM approach show generally good agreement with previous scallop assessments (Dobby et al., 2017). The absolute level of F presented here is not directly comparable with previous assessments as different age ranges have been used to define these quantities (ages 4-6 in 2017, age 4-8 in SAM assessments). At present there are insufficient data from the Clyde and Orkney assessment areas to perform analytical assessments or evaluate stock trends.

The stock recruitment plots provided for the five areas assessed using SAM show little evidence of a stock recruitment relationship and for this reason, recruitment is modelled as a random walk. One explanation for this lack of relationship is that the model estimates of SSB (biomass of individuals aged three and above) may not be a good measure of spawning potential either because a proportion of two year old individuals are also likely to be mature but are not included in the model, or because SSB does not sufficiently account for the greater reproductive output of larger individuals. Another reason could be that recruitment is largely independent of stock size (although others have observed density dependent effects; e.g. Vahl, 1982) and is driven more by external factors such as environmental conditions, which are not included in the models.

4.2. Stock status and management considerations

Although there are currently no agreed MSY biomass or fishing mortality reference points for scallop stocks around Scotland, an evaluation of stock status is provided in relation to the precautionary reference points (F_0.1 and B_pa) developed from the stock assessment outputs in this report (Table 4.1).

Table 4.1. Summary of stock status in relation to precautionary reference points.

With the exception of the North West assessment area, stock biomass in all areas with analytical assessments is estimated to be above the precautionary reference point in 2022. Most assessed areas have seen a substantial decline in landings, resulting in declining fishing mortality to below F_0.1. The exceptions to this are Shetland and North East areas where fishing mortality is estimated at just above the reference point.

Scallops around Scotland are currently managed through technical conservation measures (MLS regulations) and through measures to restrict effort (limited licenses, zonal dredge limits). Management advice is provided on the basis of qualitative precautionary considerations and the assumption of a relationship between fishing effort and fishing mortality.

Recruitment is clearly important to the fishery. In most of the assessment areas, periods of highest landings are associated with good recruitment, which in turn appear to drive upturns in SSB. Successive recruitments appear to be correlated, with high and lows evolving over four to eight years. During periods of low recruitment, there may be a need to reduce fishing mortality (resulting in reduced landings) to enable a stock to rebuild to (or be maintained at) a level that allows for MSY in future.

For the East Coast and West of Kintyre assessment areas, where both F and SSB are estimated to be within safe limits relative to provisional reference points, advice is for no increase in effort. However, given the current status of these stocks, a small increase in fishing effort is considered unlikely to be detrimental to the stocks.

For the North West, while fishing mortality is considered in safe limits, SSB is currently below the precautionary reference point, advice is for no increase in effort (such that fishing mortality remains below the F_MSY proxy) and to consider measures to safeguard the spawning stock at a level that will support MSY for future generations.

For the North East and Shetland^[2], SSB is above the provisional reference point while the stock is fished just above F_MSY proxy, and advice is for a decrease in effort.

4.3. Reference points

The lack of a clear stock recruitment relationship (as is apparent for scallops) is often assumed to preclude the calculation of target reference points based on maximum sustainable yield (MSY). However, ICES have derived MSY reference points based on an approach which uses stochastic projections to account for uncertainty in the stock recruitment model, in addition to random deviations from the model (ICES, 2016). The software which has been developed to conduct these simulations (Eqsim, part of the ‘msy’ R package) is also able to account for uncertainty in other population parameters such as weights-at-age and fishery selectivity.

This assessment report uses an alternative approach to derive fishing mortality reference points based on per-recruit analysis (ICES, 2010). ICES has advised on the use of F_MAX (fishing mortality at the maximum of the yield-per-recruit (YPR) curve) as an appropriate F_MSY proxy unless there is evidence of poor recruitment at such levels of fishing mortality. In cases where the maximum of the YPR curves is less well defined then F_0.1 (fishing mortality at which the slope of the YPR curve is 10 % of the slope at the origin) or reference points based on spawning biomass per recruit are likely to be more appropriate proxies, which is assumed to be the case for the current assessments.

The derivation of precautionary biomass reference points (used as limits rather than targets) for the assessments presented here also adopts an approach widely used by ICES. For many ICES fish stocks, B_lim (limit reference point for biomass) has been defined as the historical lowest observed spawning stock (B_loss) – the value below which recruitment is expected to be ‘impaired’ or the stock dynamics are unknown. The precautionary reference point (B_PA) is derived from this value by adjusting it to account for variability and uncertainty in the assessment (ICES, 2016).

Scallop (Placopecten magellanicus) stocks off the north east US coast are managed in relation to a target fishing mortality of 80% of F_MAX (used as a proxy for F_MSY). A proxy for B_MSY on the basis of the product of B_MAX (biomass per recruit at F_MAX) and the median number of recruits per tow from the survey is also used (SAW Invertebrate Subcommittee, 2004). The threshold for being in an ‘overfished condition’ is defined as half of B_MAX. In New Zealand, F_0.1 is used as a target fishing mortality in the major scallop (Pecten novaezelandiae) fisheries (New Zealand Government, 2011).

There are clearly a number of options to be explored for the calculation of appropriate reference points for Scottish scallop stocks. The calculation of fishing mortality reference points using the Eqsim software would be relatively straightforward given that the required inputs for the calculations are a direct output from the SAM assessment. However, testing these reference points within a management strategy evaluation framework would be a more time-consuming procedure. The further development of appropriate reference points would enable the provision of fishery advice consistent with MSY principles.

4.4. Comments on the quality of the data and assessment

The accuracy and precision of the estimates of stock status depend on the quality of both the total commercial catch-at-age data and the survey indices-at-age. The catch-at-age data are derived from landings length- and age-structured data sampled by MD staff which are then raised to total official landings data. The introduction of buyers and sellers legislation in 2006 is thought to have improved the accuracy of reported landings, although given that Scottish scallop fisheries are not regulated through TACs there is actually no incentive for fishers to underreport or misreport scallops.

Previously, the allocation of landings from statistical rectangle 40E4 between the Clyde and West of Kintyre assessment areas was known to be at odds with the apparent fishery distribution in the area. Following an investigation of the landings data, a new estimation procedure has been implemented that uses landing port information to improve the accuracy in the split of landings between the two assessment areas. This new procedure will have reduced the bias in the estimates of landings and subsequently, biases in the stock assessment results.

There are insufficient data from the Clyde and Orkney areas to perform analytic stock assessments. Clyde and Orkney have historically been less important scallop fishery areas and the unpredictable nature of these fisheries can make the acquisition of landings samples particularly difficult. The number of samples from the Clyde has increased in recent years and the resulting data, when combined with the developing survey time series (first conducted in 2019), could potentially form the basis of an assessment in future years. As a first step, approaches to working up the survey data will be explored in the near future. The roll out of REM across the scallop dredge fleet may also provide data which could in future be utilised in the assessment process for both data limited and data rich areas (e.g. high resolution effort data to develop landings-per-unit effort indicators).

In the scallop areas for which MD have conducted analytical assessments, sampling levels have historically been relatively good for the west of Scotland areas (and Shetland) and poorer for the North Sea assessment areas. In recent years however, there have been significant reductions in the number of samples across all assessment areas (except Shetland) and although a single year with poor sampling levels may not significantly affect the conclusions of the assessment, continued poor sampling levels are likely to result in less robust results. In order to account for the variation in sampling levels over time, the number of catch samples per year is used to weight the data such that the catch data from those years with a greater number of samples is given a greater weight in the assessment, hence reducing the impact of particularly noisy data from years with poor sampling. Work is ongoing within MD to explore ways of improving future sampling levels including working with local scallop processors.

The current analytical assessments rely on age-structured input data. In some areas (particularly the West of Kintyre) these data appear more noisy which could be due to age reading errors although equally it could reflect heterogeneity of the stock or fishery in the area. Within MD, age-reading training is conducted on a regular basis, however, ICES have identified the need for internationally agreed ageing protocols along with additional quality control and assurance to validate age and growth estimates and improve consistency among readers. A recently established ICES workshop has been tasked with progressing some of these issues (Workshop on Scallop Ageing; ICES, 2023b).

The survey data are an integral component of the stock assessments. The surveys show reasonably good coverage of the fished areas according to scallop dredge VMS data with the exception of the West of Kintyre where there are a number of areas with apparently high fishing effort which are not surveyed (Dobby et al. 2017). The density of stations is greatest in Shetland although in some years no survey (or only a partial survey) has been completed due to poor weather. It is not clear whether such a high density of stations is required to retain a particular level of precision in the survey abundance index estimates. In recent years a number of stations to the west of Scotland have not been surveyed due to the presence of newly designated marine protected areas (MPAs) and in the North Sea due to renewable energy developments. Previously an analysis of the historical survey data suggested that the survey index was relatively insensitive to the inclusion/exclusion of a small number of survey stations. Many fish stock assessments now make use of modelled survey indices (e.g. Northern Shelf haddock and cod; ICES 2024a) and these can provide a more reliable index with variable survey coverage and also provide a statistically sound approach to the derivation of survey confidence intervals which can be included as assessment input data.

The analytic stock assessments presented here utilize SAM which is a state-space stock assessment model (Nielsen and Berg, 2014; Berg and Nielsen, 2016) similar to Time Series Analysis (TSA) which was used for the previous stock assessments (Dobby et al, 2017). The new approach has substantial flexibility in terms of model configuration and input data, allowing for both process and observation error and providing model outputs with estimates of precision. SAM has the advantage (over TSA) of being a widely used stock assessment method with a substantial network of user support, and in addition utilizes an online platform on which to run assessments and make data and results publicly available (stockassessment.org).

Comparisons of estimated historical stock trends between the two assessments indicate that the results are fairly robust to the choice of assessment method. However, of the five analytical assessments presented, two show substantial retrospective revisions (North West and Shetland) with Mohn’s rho estimates of average bias falling out with the ICES guidelines (ICES, 2020). In such situations ICES would likely consider either scaling the assessment results to account for the bias before using them as the basis of numerical catch advice, or potentially making some additional precautionary considerations. For these scallop stocks, advice is qualitative (rather than numerical) and the conclusions in terms of biomass stock status are likely to be robust to the retrospective bias in this case (given the direction of bias and the biomass estimates relative to the reference point).

The current stock assessments provide an indication of stock status (and dynamics) in the assessment areas as currently defined. These areas were, however, based on the characteristics of fisheries in the past rather than on the basis of evidence of discrete populations. The population structure of Scottish scallop stocks is not well understood. Scallops are sedentary in nature and only able to swim limited distances. Larvae, however, inhabit the water column for three weeks or more, during which time they may drift a substantial distance (dependent on water circulation, tides and wind driven currents) from the parent population before settling to the sea bed. There appear to be some similarities in interannual fluctuations in recruitment between assessment areas (e.g. North East and East Coast) which may indicate that there are linkages between some of these areas at pre-recruitment stages with similar trends in survival to age of recruitment. There is potential for population linkage across substantial distances. Habitats suitable for scallops are patchily distributed and some patches of adult population may provide a source of larvae for others. A number of PhD projects on genetic connectivity and larval dispersal modelling around Scotland are currently on-going (ICES, 2024b). These will provide an improved understanding of the nature and extent of connectivity between scallop populations around Scotland and further afield.