Crab and Lobster Fisheries in Scotland: Results of Stock Assessments 2009-2012: Scottish Marine and Freshwater Science Vol 7 No 9

4. General Discussion

4.1. Landings

Brown crab remains the most important species, in terms of landed weight, in the crab and lobster fisheries around Scotland. Landings from Orkney, East Coast and Ullapool increased, whereas those from the offshore areas of Papa and Sule seem to have stabilized in the last four years. The latter may be related to market prices, which have not increased substantially in recent years, and reduced effort rather than any change in the abundance of the stock offshore. Total landings of velvet crab have remained stable in recent years. Increases in reported landings from traditionally less important areas such as the East Coast, South East and Shetland have coincided with decreases in Orkney and South Minch. The spatial distribution of the velvet crab fishery is similar to previous years and the fishery continues to take place in inshore areas. Lobster landings, although still much lower than those of brown and velvet crabs, have more than doubled since 2001, with the East Coast and South East areas making the major contribution to this increase.

Landings of all three species are thought to have been under recorded prior to 2006 before the introduction of the "buyers and sellers" legislation. Some major increases in landings evident in the mid 2000s are more likely to be explained by improved reporting than by changes in the abundance of stocks. There is little information on changes in fishing effort over the past 30 years, but it is likely that technological advancement and mechanisation of fishing and processing have allowed the crab and lobster fishery to expand and effort to increase. The emergence of European markets, and more recently eastern markets in China, combined with the ability to transport live animals has increased the demand, particularly for brown crab. However, this has not been accompanied by an increase in crab prices. Market and supply issues contributed to the setting up of a European project ACRUNET (A Transnational Approach to Competitiveness and Innovation in the Brown Crab Industry) in 2012. The project comprises several work packages with the collective aim of securing the economic and social viability and sustainability of the European brown crab industry.

4.2. Stock Assessment

4.2.1. Brown Crab

Brown crab sampling levels were relatively high in most areas where landings are important. In the northwest of Scotland, female crabs usually make the largest contribution to the landings. This suggests that females are exposed to a higher exploitation than males in some of the offshore grounds such as the Hebrides or Sule and, in general, this is supported by the LCA results. Evidence from tagging studies suggests that females migrate from deeper offshore grounds, where they remain for most of the year, to inshore areas where they moult and mate with males (Edwards, 1979; Jones et al., 2010). Large females also predominate in the landings from offshore areas to the north of Ireland (Tully et al., 2006), and in the fisheries off the east coast of England and in the English Channel, aggregations of ovigerous females have been observed (e.g. Howard, 1982). The average size of the largest 20% of individuals landed show an increase in the East Coast, Hebrides, North Coast, Orkney and Shetland, in particular for male stocks. This could be indicative of an increase in the proportion of large individuals in the stock, and possibly a reduction in fishing mortality. It could, however, also be a reflection of changes in fishing or discard practices. Without further fishery information it is difficult to conclude which is more likely.

Assessments based on the 2009-2012 data (using LCA) were carried out for nine out of the twelve assessment areas. Stocks in most of the areas remain growth overfished to some extent, especially among the females. For those stocks fished above the F _MSY proxy, a long term reduction in effort would increase biomass-per-recruit. The East Coast and South East male brown crab yields in particular, would benefit from an effort reduction as they appear to be significantly growth overfished. The mean sizes of brown crabs in the South East and East Coast assessment areas were smaller than the national average, as found in previous assessments (e.g., Kinnear, 1988; Mill et al., 2009). It should be noted that the growth parameters used in the assessments are the same as those applied in the other areas and this may not be appropriate if individuals in the East and South East are slower growing and reach smaller maximum sizes. Use of inappropriate growth parameters in the LCA could lead to overestimation of F and the conclusion that the stocks are more heavily fished than they actually are. This is further discussed in section 4.3. Assessment results in the North Coast (both sexes), and male stocks in the Hebrides, Sule and Shetland, show an improvement compared to those reported previously (Mill et al., 2009; Mesquita et al., 2011), with crab now fished below or around the F _MSY proxy.

4.2.2. Velvet Crab

For velvet crabs, males were more common than females in the landings sampled, and were generally slightly larger. Mean sizes in landings are quite variable from year to year and in areas where data collection is more sporadic. This may reflect changes in the sampling effort over the time series, rather than changes in the population. The average size of largest individuals in the sampled landings showed an upwards trend in the Hebrides. If the landings LFDs are representative of the population size structure and there have been no recent changes in the exploitation pattern, this would indicate that the fishing mortality may have declined over the time series.

LCA results showed that velvet crab stocks in most of the areas were growth overfished to some extent in 2009-2012. Clyde stocks and female stocks in Orkney and South Minch would benefit from an effort reduction as they appear to be significantly growth overfished. For both Orkney and Hebrides stocks, fishing mortality showed a decreasing trend from the 2002-2005 assessment estimate (Mill et al., 2009), although in 2009-2012 Orkney remained above the F _MSY proxy while Hebrides velvet crabs were fished below (males) or just over (females) F _MSY.

4.2.3. Lobster

Lobster market sampling data from 2009-2012 suggest that male and female lobsters were generally landed in equal proportions. For most areas, the mean size of the largest individuals has remained stable in recent years. However, in the Hebrides and South Minch mean size in landings of larger animals has decreased, which may indicate an increase in the long term fishing mortality although this could also be affected by changes in the fishing practices. The recent decline observed in the Hebrides could also be related to sampling noise as the number of sampling trips declined markedly in 2012.

LCA results showed that stocks in most of the areas remained growth overfished to some extent in 2009-2012, especially among the males (with the exception of Shetland), which would benefit from an effort reduction to achieve a higher yield and biomass in the long term. Females in Papa, Hebrides and Orkney were fished below or at the F _MSY proxy while in other areas F is over F _MSY. There are differences in mean size in the sampled landings between areas, with lobsters from the North Coast and Hebrides being significantly larger than those from the South East and East Coast areas. It should be noted that the growth parameters assumed are the same for all areas, except Shetland. This may not be appropriate if individuals in some areas such the East Coast and South East are slower growing and attain smaller maximum sizes. The use of inappropriate growth parameters in the LCA could overestimate the magnitude of growth overfishing. This is further discussed in section 4.3.

4.3. Quality of the Assessment and Data

4.3.1. Landings Data

From the range of stock assessment tools available, LCA is one of the least data intensive and LCA and yield-per-recruit models are often the preferred methods, or the only methods available, for assessing data-poor shellfish stocks. A major assumption of the LCA is that the landings length frequency distribution is representative of the fishery removals from a single cohort of individuals throughout its life. However, since the length frequencies are derived from a single year of sampling, rather than from a single cohort, then this assumption is only true if the population is in a steady state or at equilibrium, i.e. that recruitment and exploitation rate are constant. Landings data from most of the Scottish assessment areas tend to fluctuate, which may reflect year to year variation in recruitment or fishing effort. An average of the length frequency data (2009-2012) was used in order to limit the effects of these variations. However, any systematic changes in exploitation rate or recruitment over this four year period could still result in biased estimates of fishing mortality.

The length and sex composition of the landings data is considered to be well sampled in most areas where landings are important. However, in some areas such as Papa (brown crab and lobster), North Coast (brown crab) and South Minch (lobster), size data are sparser, and sampling levels remain relatively low with few sampling trips taking place. Length frequencies derived for these areas are often similar to those in other adjacent areas and therefore assumed to be representative, despite the low sampling levels. The effect of this assumption on LCAs has not been investigated and the results of these assessments should be interpreted with caution.

4.3.2. Biological Parameters

LCA is frequently used for assessing crustacean species for which ageing techniques are not yet fully developed. Discontinuous growth within a cohort of animals (growth increments and frequency of moults) results in growth rates varying between individuals within a stock. This can make it difficult to track cohorts through length frequency data and hence assessment methods which translate between size-structured and age-structured data have not been widely applied to crustaceans (e.g. Sheehy et al., 1996; Sheehy and Prior, 2008; Kilada, 2012).

In addition to landings length frequency distribution data, LCA also requires estimates of other biological parameters, including von Bertalanffy growth parameters and natural mortality. LCA is very sensitive to these parameters and the choice of input parameters may critically influence the results obtained (Lai and Gallucci, 1988; Jones, 1990), such as the perception of the state of the stock, in terms of the position of the current exploitation rate in relation to the F _MSY proxy. Natural mortality ( M), for example, has a marked effect on the shape of the relative yield per recruit curve. Using lower values for M results in a more pessimistic stock assessment, with current fishing mortality estimated to be higher in relation to the F _MSY proxy (or vice-versa). The values of the von Bertalanffy growth parameters ( K and L _∞) also affect the shape of the relative yield per recruit curve and estimation of the value of F in relation to the F _MSY proxy. Using growth model parameters that result in growth rates that exceed the true growth rate (i.e. using values of K and L _∞which are too large), results in the current exploitation rate (F) being over-estimated in LCA and could lead to the erroneous conclusion that a stock is growth-overfished (and vice-versa). Currently, the same biological parameters, those which were used in previous assessments (Chapman, 1994; Mill et al., 2009; Mesquita et al., 2011), are applied across all regions except Shetland. Studies of velvet crabs in Shetland (Tallack, 2002; Mouat et al., 2006) provided much higher estimates of M (0.576 compared to 0.1 elsewhere) and a higher value of K (0.463 compared to 0.1 elsewhere). The high value of K estimated for Shetland velvet crabs implies a very fast growth rate for young crabs (for example a 30 mm male crab would be expected to grow to nearly 60 mm carapace length in a single year). This seems unrealistic and merits closer scrutiny of the original data and methods used to provide the growth parameters for this stock. The marked difference in parameters is sufficient to explain the disparity in the results obtained for the velvet crab assessments in Shetland when compared to other stocks with relatively similar length frequency distributions. Male velvets at Clyde, East Coast and South East are estimated to be growth overfished, whereas in Shetland they are estimated to be fished below the F _MSY proxy, despite the estimated F in Shetland being much larger than those in other areas. Owing to the uncertainty around appropriate input parameter values, care is required in drawing firm conclusions regarding the status of the various velvet crab stocks. To progress this discussion it would be worthwhile holding a joint NAFC Shetland/ MSS crab working group in the near future.

Differences in size composition across areas, particularly the relatively small size of brown crab and lobster in the South East and East Coast compared to the north and west, suggest that area specific parameter values may be more appropriate and it is possible that the extent of growth overfishing of brown crab and lobster in the East Coast and South East may be overestimated. A re-evaluation of the parameters used in LCA in these areas would be extremely useful. It would require a large scale tagging project using tags that could be reliably retained on moult, with seasonal measurements of length and weight.

4.3.3. Reference Points

Given that LCA provides only long term equilibrium predictions, it is usually advisable to complement the outputs with additional data which can provide information on trends in abundance, typically catch per unit effort data or exploitation rate. Effort data in terms of numbers of creels fished are not currently available, precluding calculation of catch per unit effort in these fisheries. In an attempt to gain additional information on variation of fishing mortality from the data which are available, trends in the mean size of largest 20% of individuals (Mesquita et al., 2011) were explored.

In several of the well sampled areas, trends that could be associated with changing exploitation rate or recruitment are apparent. For example, male brown crab in the East Coast and Shetland areas and velvet crab in the Hebrides area show an increase in the average size of the largest individuals over the time series which could be related to reduced fishing mortality. These areas also show sporadic occurrences of large numbers of small individuals in the length frequency data, which could be due to increased recruitment in these years. On the other hand, brown crab in the South East and velvet crab in the East Coast exhibit a decreasing trend in the mean size of larger individuals, particularly in the last 10 years. This could indicate an increase in the fishing mortality but also changes in the fishing patterns (e.g. changes in discard practices) of the fishery. In other areas such as South Minch (brown crab), South East (lobster) and Orkney (velvets and lobster), the mean sizes and length frequency data appear relatively stable suggesting a stable stock and fishery. In other cases, however, the data show significant inter-annual variability (although no trend) which could be due to low and erratic sampling. Trends could also reflect changes in fishing practices (e.g., retention size) as well as changes in actual population size structure. This emphasizes the need for improved sampling and better information on fishing activity and fishers' behaviour to develop robust size based indicators for assessment purposes.

The conclusions in this report are all based on estimates of fishing mortality in relation to a reference point for each stock to infer whether or not a stock is growth overfished. Although LCA and yield-per-recruit analysis give an indication of current F relative to the fishing mortality required to optimize yield (from a particular cohort), it provides no indication of whether or not a stock is recruitment overfished (i.e. whether fishing is compromising recruitment). F _MSY, the fishing mortality which gives the maximum sustainable yield (high long term yield with low risk of stock depletion), is difficult to estimate. In addition to a yield-per-recruit curve, calculating F _MSY requires good estimates of recruitment and spawning stock biomass, which are not available for Scottish crab and lobster stocks. In cases where F _MSY cannot be estimated directly, proxy values based on per recruit analysis are often used. ICES advises that in cases where the peak in the yield per recruit curve is well defined and there is no evidence of poor recruitment at this level of fishing mortality, then F _MAX may be an appropriate proxy for F _MSY ( ICES, 2010). In cases where the maximum is less well defined and the curve is more flat topped, then F _0.1 is likely to be a more appropriate proxy (Jennings et al., 2001). Another potential reference point is F _30%SpR which is defined as the fishing rate which results in combined SPR equal to 30% of the un-fished level. F _0.1 is usually the most conservative reference point while F _MAX is generally above F _30%SpR, depending of the relative shape of the YPR and BPR curves. F _MSY proxies for Nephrops stocks assessed by ICES have been selected from these three candidate reference points (F _0.1, F _35%SpR or F _MAX) for each stock independently according to the perception of stock resilience, typical population density, biological knowledge and the nature of the fishery (e.g. ICES, 2015). Most crab and lobster fisheries have been in existence for several decades with little evidence of between stock differences in resilience. Therefore, despite some stocks showing very flat topped YPR curves (which might suggest F _0.1 as the most appropriate proxy for F _MSY), F _MAX was selected as a proxy for F _MSY. For the purposes of consistency in this report, all discussion relates to the F _MSY proxy (F _MAX) although F _0.1 and F _30%SpR values are also shown in Figures 15- 20. The reference points remain preliminary and may be revised in the future as further data become available.

In most areas around Scotland, the crab and lobster stocks are being fished at levels which result in yield per recruit values not far below the maximum. In some cases, the estimated fishing mortality is substantially above the F _MSY proxy, making it more likely that these stocks (e.g. South East male brown crab and most male lobster stocks) are recruitment overfished as well as growth overfished. It should be noted however, that, so far, lobster stocks have not showed signals of systematic changes in sex ratio which has been associated with overfishing in other lobster species.

4.4. Data Gaps and Future Research Priorities

From the discussion above it is clear that there are a number of areas where research or additional data collection would improve Scotland's crab and lobster stock assessments.

4.4.1. Fishing Effort

Currently, no useful measures of creel fishing effort are available from official log sheets which precludes the use of LPUE data as an indicator of abundance for the crab and lobster stocks around Scotland. Fishing effort for most finfish species can be estimated as fishing time using days absent from port and vessel power but these are not particularly useful measures of effort in creel fisheries. The number of creels used or hauled when fishing for crabs and lobsters would be a much more useful measure of effort in the fishery, and could be used to calculate landings-per-unit-effort. Shetland is currently the only area for which effort data are available and routinely collected. The Shetland Regulating Order requires licensed fishers to return logbook information to the SSMO, detailing the catch location (at the 5 nmi scale) and details of the number of creels or pots fished. This data collection method would appear transferable to other areas and would allow LPUE to be calculated in an appropriate manner (Leslie et al., 2010). A recent EU project (Anon, 2010) investigated the use of voluntary logbooks with the aim of improving brown crab data collection and concluded that although data collected as part of such schemes were generally 'fit for purpose', uptake of such schemes is typically low. In addition, it was concluded that for inshore vessels whose catch and effort are likely to be highly variable, a high level of industry uptake would be required to ensure data precision and accuracy and that the data were representative of a particular geographic region.

VMS (vessel monitoring system) data have become available for larger vessels (over 15 m from 2008 and over 12 m from 2012), and could potentially be integrated with logbook landings information to obtain indicators of LPUE for the offshore fleets. However, these monitoring tools do not cover the majority of the inshore fleet which is mostly composed of smaller vessels (under 10 m). Scotland's Inshore Fisheries Groups ( IFGs) may provide the impetus and the means to improve the knowledge base for the management of stocks locally. Data collection would, however, need to be coordinated and maintained, to build up useful time series, and in the meantime the need to improve the information collected on a national basis for both inshore and offshore fisheries is likely to remain.

4.4.2. Size-based Indicators

ICES has suggested a multiple indicator based approach (including LPUE, size-based indicators and recruitment indices) as a potential way forward in the provision of advice on stock status for crab stocks ( ICES, 2009). More recently, ICES WKLIFE ( ICES, 2014b) discussed approaches to develop reference points based on length frequency information and biological parameters. Size-based indicators have been investigated in this report, and in some cases it was possible to relate variations of the mean size of largest animals to trends in fishing mortality. However, market sampling in some areas remains low, and discards in crab and lobster fisheries are sampled only on an irregular basis. More regular sampling to obtain information on catches of undersized animals could provide a better indication of inter-annual variation in recruitment. By-catch data collected on MSS scallop surveys may potentially provide information on recruitment variations in areas where the scallop and the crab and lobster distributions overlap. Further discard studies are also required to obtain estimates of discard survival and to help understand more fully the reasons for discarding.

4.4.3. Population Structure

The population structure of crab and lobster stocks around Scotland, and the rest of the UK, is not well understood. The current assessment areas are empirical, based largely on past fishing patterns. Brown crab are known to undertake extensive seasonal migrations in some areas while in contrast velvet crabs and lobsters appear to make relatively limited movements. MSS recently conducted a tagging study of brown crab to the north of Scotland (Jones et al., 2010). The results suggest linkage between inshore and offshore crab stocks to the north and west of Scotland. Fishermen support the idea that crabs migrate between (and across) the 'windsock' and inshore grounds around Orkney, although there is only a limited fishery in the area in between. Large catches of female crab have also been reported on the shelf edge at depths greater than 200 m. Work being undertaken in both Shetland and Orkney, should provide further evidence regarding the structure of brown crab stocks to the north of Scotland. Ideally, such studies should be followed up by population genetics/morphology studies and consideration of larval dispersal. It has been suggested, for example, that brown crab populations in the Irish Sea may be closely linked to the larger populations on the Malin shelf, which are contiguous with the Hebrides and South Minch in western Scotland (Tully et al., 2006; ICES, 2007). The current brown crab assessment areas will be reviewed by ICES WGCRAB in the future.

4.4.4. Other Factors

The interpretation of trends in indicators derived from fishery dependent data would be helped by improved understanding of the economic and environmental factors that influence fishers' decision making with respect to fishing location and target species. A component of the Lot 1 EU project (Anon, 2010) involved conducting questionnaires and interviews to establish the main factors in fishers' decision making. As well as providing information on historical changes in fishing practices, the interview responses suggested that in recent years the Scottish brown crab fishery has been influenced more by the market than by management measures. Many fishers acknowledged that low brown crab market prices in the late 2000s resulted in the species being targeted to a lesser extent than usual, which could explain the reduction in landings in some areas. Additional information on factors affecting catchability such as bait type, creel density and soak time could also be collected by engaging with fishers and industry.

Acknowledgements

The authors would like to express their thanks to the various Marine Scotland Science and Marine Scotland Compliance staff who collected and collated the data used in this report. We would also like to thank the NAFC Marine Centre and the Shetland Shellfish Management Organisation for providing size frequency data for the Shetland area.