Collection of Data to Inform the Implementation of a Discards Ban: Scottish Marine and Freshwater Science Vol 7 No 12

5 Discussion

The primary aims of this project were to collect quantitative information on the amount and nature of the fish currently being discarded in the mixed whitefish fishery around Shetland, and to implement and evaluate a discard self-sampling programme in this fishery as a means of increasing the quantity of discards data from Scottish whitefish fisheries. Although direct verification was not possible, the results of the analysis of the fishermen's discard samples were compared to other available observer data to evaluate their reliability.

The following discussion first considers the information that was collected through the self-sampling programme on nature and scale of discarding, then considers self-sampling as a means of collecting such data, and finally considers some of the potential implications of the discard ban for fishermen.

5.1 The Nature and Scale of Discarding

5.1.1 Composition of Discards

With a few exceptions, the composition of the fishermen's discard samples was broadly similar between the different fishing gears. One of the most obvious differences was that while saithe was one of the dominant species in the discards from the trawl net vessels (single and twin) it was almost entirely absent from the discards of the seine net fishing vessels. This probably reflects the fact that seine net vessels tend to catch less saithe because it is generally less common on seine net fishing grounds. Anecdotal information from fishermen indicates that saithe tend to be associated more with areas of harder (rougher) sea-bed where seine net vessels cannot operate. In 2014 saithe accounted for only 3% of the total landings of the seine net vessels involved in this trial, compared to 12% for the trawlers.

The variability in discard composition between the fishermen's discard samples was relatively small, both overall and for individual fishing gear types. This suggests that the composition of discards does not vary widely over time or between fishing vessels. It also gives some confidence in the reliability of the self-sampling of discards by fishermen (in that inconsistencies in the sampling might result in large variability between the samples).

The composition of the fishermen's discard samples broadly agreed with the composition of the discards reported through the SFA's discard tally book scheme (which included several of the same vessels). The same species dominated the discards in both cases and their relative proportions matched closely overall (despite some differences in detail).

5.1.2 Discard Rates

The overall discard rate of 42% by the vessels sampled in this study was very close to the figure of 40% reported through the SFA's discards tally book scheme. The discard rates varied for the different fishing gears, being highest for the seine net vessels and lowest for the single-trawl and camera-equipped twin trawl vessels, but again matched closely with the rates derived from the tally books. The variability in the discard rates calculated for the individual samples was relatively small.

Hake had the highest estimated individual discard rate, with some 80% of the catch being discarded overall. Rays, saithe, ling and cod also had relatively high estimated discard rates (above 20%). Other commercially important species had lower discard rates. (Note: these estimated discard rates for individual species are based on a number of assumptions, in particular that the sampled catches are representative of the catches of the Shetland whitefish fleet, and so should be treated with caution).

A number of non-target species also had very high estimated discard rates (100% in many cases). These included species of little or no commercial value (such as dogfish or gurnards), species that cannot be landed (such as common skate), but also included several pelagic species (including herring and mackerel). Although the quantities of the pelagic species being discarded were not large (perhaps a few hundred tonnes per year) they will be covered by the discards ban. However, whitefish fishing vessels generally lack the quota needed to land these species.

There was some variability in the estimated discard rates of individual species between the different fishing gear types, although there was broad agreement in the overall pattern.

The discard rates estimated for individual species from the fishermen's discard samples matched well with the those determined from the SFA's discard tally books, and with those measured by observers (from the Data Limited Stock project). There was also good agreement between the discard samples and the discard tally books for the individual fishing gear types (the agreement with the observer data for individual gears was not statistically significant). Again, this agreement gives some confidence in the reliability of the self-sampling of discards by fishermen, and of the tally book scheme (in that inconsistencies in either scheme would be expected to result in differences between the results).

The general pattern of the discard rates estimated from the fishermen's discard samples for cod, haddock, hake, monk, saithe and whiting was also similar to published discard rates (Needle et al., 2014; Figure 19), especially those collected through two other Scottish observer programmes. Differences between these data (higher discard sample rates for hake and lower rates for haddock and whiting) might be due to differences in the geographical areas covered by the different programmes. While the discard samples in this study were collected from vessels working in the northern North Sea, the Scottish observer programmes covered vessels fishing throughout the North Sea. The match between the discard sample rates and the published CCQS (camera) rates was less good, but no worse than difference between the CCQS and published observer programme rates.

Figure 19: Comparison of discard rates estimated from the fishermen's discard samples (all gears) with published rates (from Needle et al., 2014). MSS Observers = Marine Scotland Science observer programme; SFF Observers = Scottish Fishermen's Federation observer programme; CCQS (cameras) = analysis of CCTV footage from vessels equipped with cameras under the Cod Catch Quota Scheme (which are prohibited from discarding cod; see footnote on p. 10). MSS, SFF and CCQS data are overall means for Scottish whitefish vessels fishing in the North Sea during Q4 of 2012 and Q1 - 3 of 2013. Error bars on the Discard Sample estimates show the standard error of the mean (not available for the published data.

5.1.3 Quantities Discarded

The quantity of any particular species discarded is a function of both the discard rate and the quantity caught. Thus, even if a species has a high discard rate the amount actually discarded may not be large if the quantity caught is small. Conversely, a species with a low discard rate may actually be discarded in large quantities if it is caught in large quantities. Overall, the data collected from the fishermen's discard samples suggest that Shetland whitefish fishing vessels may have discarded a total of some 6,600 tonnes of fish in 2014, compared to landings of about 13,300 tonnes. (Given that this estimate is based on a number of assumptions it should be treated with caution).

5.1.4 Size Distributions

The vast majority of the fish in the fishermen's discard samples were above the minimum landing size. For most species more than 95% of the measured fish exceeded the minimum landing size, for some (such as hake) it was 100%. This contrasts with the conclusion by Heath & Cook (2015) that the majority (60-65%) of the fish discarded in mixed demersal fisheries in the North Sea were below the minimum landing size. However, their estimate is based on modelling of historical data collected over the period from 1978 to 2011 for the whole North Sea and so may not be comparable with the results of this study.

Several broad patterns were evident in the size distributions of the discarded fish. For some species, such as haddock and whiting, the size distribution tended to be fairly narrow, with most discarded fish close to the minimum landing size. This suggests that discarding of these species primarily resulted from the grading of fish and the discarding of fish close to the minimum landing size (lack of quota is not generally an issue for species like haddock and whiting).

For some other species, such as cod, hake and saithe, the size distribution of the discarded fish is much wider, with most (in some cases all) of the discarded fish above the minimum landing size. This suggests that these species are primarily being discarded due to a lack of quota (over the course of the year), rather than size limits.

A third pattern was evident for plaice, with the majority of the discarded fish above the minimum landing size although plenty of quota was available. This is believed to reflect discarding of sizes for which there is little market demand.

There was some agreement between the size distributions of the fish in the fishermen's discard samples and those measured by observers under the Data Limited Stock project, with the exception of hake where there was a consistent difference. Even where there were statistically significant differences the magnitude of the differences tended not to be large (< 10 cm for hake; perhaps a few cm for ling).

5.1.5 Discard Data: Conclusions

The results obtained from this analysis of fishermen's discards samples should be treated with some caution: As is discussed below (Section 5.2.1), the majority of the samples were returned by a small number of vessels and the randomised sampling protocol was not followed in all cases.

Nevertheless, the discard self-sampling programme has provided a substantial amount of new information on the nature and scale of discarding by vessels in the Scottish mixed whitefish fishery in the northern North Sea. The results obtained were internally consistent (variability between samples was relatively small) and agreed well with those from other sources (in particular the SFA's discard tally book scheme). They also agreed well with published discard rates from other Scottish observer programmes (Needle et al., 2014).

Thus, it is suggested that the data obtained through the self-sampling of discards in this project can be regarded with some confidence, although the absence of observers in this project precluded any direct verification.

The discard self-sampling has confirmed the high levels of discarding of hake, saithe and cod in the northerly component of the Scottish mixed whitefish fishery and has shown that the vast majority of the discards of these species are of marketable size. These species arguably represent the biggest discard 'problem' for the Shetland whitefish fleet, due to the relatively large quantities being discarded and the difficulty of reducing their catches. (Hake, saithe and cod are all relatively large fish so reducing their catches without losing smaller, valuable species with lower discard rates, such as haddock and whiting is a particular challenge).

A related point that can be taken from these results is that very few of the fish being discarded are 'undersized' (that is, below the minimum landing size). Overall, only 8% (by weight) of the discarded fish were below the minimum landing size (and if ling are excluded the figure is less than 4%). The vast majority of the commercial species in the fishermen's discard samples could have been landed and sold if the vessels had had sufficient quota to allow this (and if there was sufficient market demand).

The discard sampling has also highlighted relatively high discard rates for some other species, both commercial and non-commercial. These included rays, gurnards and lesser-spotted dogfish, where discarding was probably mainly a result of lack of market demand. The discarded rays tended to be relatively small and there is little market demand for small rays. Similarly, although gurnards are sometimes landed in small quantities anecdotal information suggests that there is not a strong market demand for them, again especially for small sizes (although the gurnards were not measured in this study most of those in the fishermen's discard samples were relatively small).

A final point worth noting from the fishermen's discard samples is that pelagic species such as herring, mackerel, horse mackerel and blue whiting are regularly being caught and discarded by whitefish vessels. The quantities involved are not large, amounting to perhaps 200 to 300 tonnes for the entire Shetland whitefish fleet in 2014, but they pose a further problem for whitefish vessels which typically do not have quota for pelagic species.

5.2 Evaluation of Self-Sampling

The result of this study have shown that self-sampling - where fishermen themselves collect samples of their catches for analysis ashore - offers a potential alternative to the use of on-board observers for the collection of fisheries data from commercial fishing vessels.

Although the use of observers would tend to be a preferred option, as in this case they are not always available. In their absence, self-sampling provided a potential means of collecting some data on the nature and scale of discarding by whitefish vessels in the waters around Shetland.

Both the use of observers and self-sampling have other potential advantages and disadvantages, some of which are summarised in Table 13. Amongst other things, it might be said that while observers allow for the controlled, intensive sampling of a small number of vessels, self-sampling potentially allows for the less intensive (but less controlled) sampling of a larger number of vessels.

A key question is whether the data collected through self-sampling is reliable and dependable; whether the advantages outweigh disadvantages.

Although the absence of observers in this trial precluded direct verification of the data collected through self-sampling, analyses indicated that these data agreed closely with equivalent data from other sources (tally books and observers engaged in a separate project). This suggests that self-sampling of catches by fishermen can be a credible means of collecting fisheries data and can provide useful information.

As discussed above, self-sampling should probably not be viewed as a substitute for scientific observers, but rather as a potential alternative when observers are not available (or not available in sufficient numbers). An optimum sampling strategy might be a combination of the two techniques; with self-sampling allowing simultaneous coverage of a number of vessels and observers providing verification and additional data.

Table 13: A summary of some of the potential advantages and disadvantages of using observers on-board commercial fishing vessels, as opposed to self-sampling by fishermen as a means of collecting discard information.

Observers	Fishermen's Self-Sampling
High level of control over sampling protocol.	Low level of control over sampling protocol.
Requires sea-going staff (capable of working at-sea on board commercial fishing vessels).	Does not require sea-going staff.
Observer can only sample one vessel at a time.	Same staff can process samples from multiple vessels.
Limit on the amount of time that an observer can spend at sea / number of vessels that can be covered.	Limited only by the number of vessels willing to undertake sampling, and the availability of onshore staff.
Observers are occupied full-time.	Staff not required full-time.
Observer can sample all (or most) tows made during a trip.	Samples can probably only be collected from a few of the tows made during a trip.
Observer can analyse all (or a high proportion) of a catch.	Only a sample of the catch is available.
Observers' priority is to collect samples / data.	Fishermen have other priorities that may interfere with sample collection.

5.2.1 Practical Aspects of Self-Sampling

Self-sampling ultimately depends on the cooperation and assistance of the fishermen who are asked to undertake the sampling. The experience of this study showed that fishermen's responses could be 'patchy'. Some skippers and crews returned samples on a regular basis over an extended period of time, but others seemed to 'lose interest' after returning some samples. The reasons for the failure of some crews to return samples on a regular basis were unclear; no crew ever directly refused to provide samples or expressed any unwillingness to do so. Reasons given failing to return samples were generally either that they had 'forgotten' or been 'too busy'.

It needs to be borne in mind that for commercial fishermen sample collection will not be their primary priority during fishing trips. Nevertheless, if self-sampling is to be successful it will be necessary to find ways of maintaining fishermen's commitment to the sampling programme. One possible option might be more proactive engagement with the fishermen on a regular basis; for example, contacting them directly and asking them to collect a sample from their next tow. Another option might be to offer some form of recompense (such as a small payment, to fishermen for each sample landed).

It was notable that the majority of the samples returned did not come from the catch indicated by the random numbers on the sample record sheets. This could indicate a bias by fishermen in the selection of the catches sampled, but the general agreement between the results derived from the analysis of the discard samples and those derived from other sources (including observers and tally-books) provides no evidence that this was the case. (Although most samples did not come from the randomly indicated catch that does not necessarily mean that the sampling was 'non-random', in the sense that the fishermen were deliberately selecting which catches to sample, or not sample). Other than bias, possible reasons for the fishermen's failure to sample the randomly indicated catch could include practical reasons or a lack of awareness of the importance of the random numbers.

The sampling protocol, and how best to ensure that samples are collected randomly (so far as is reasonably practical), would need to be addressed in any future self-sampling programme.

The experience of this study has also highlighted some practical issues with the shore-based reception and analysis of samples that need to be considered: Firstly, it is desirable to have some means of monitoring or checking landings for samples. Whilst the fishermen in this study did usually provide notice when they intended to land their samples, or after they had done so, they did always do so. The other main point that needs to be considered is that staff need to be available, often at short notice, to deal with the samples when they area landed. Furthermore, these staff need to have access to necessary equipment such as vehicles. Finally an appropriate means is required for disposing of the material from the discard samples.

5.3 Implications of the Discards Ban for Fishermen

5.3.1 Costs

A common concern for fishermen is that under the discard ban they will be required to handle, box, store and land (and perhaps dispose of) substantial quantities of unmarketable fish. As well as the extra labour involved this could impose additional costs on fishermen in the form of charges for additional boxes, ice, etc.

The results of this analysis, however, suggest that the vast majority of the fish being discarded, of the species that will be covered by the discards ban, are potentially marketable. That is, they are of a size that could legally be landed and sold. So while fishermen might have to handle, box, ice and land more fish under the discard ban, they could expect a commercial return for that fish (all else being equal). This assumes, firstly, that fishermen are allowed to land the extra fish (see below), and that the markets can absorb the extra fish without there being a substantial adverse effect on prices.

Overall, it was estimated that the quantity of unmarketable fish discarded by the Shetland whitefish fleet in 2014 amounted to approximately 5% of the total quantity caught. The total cost of landing and disposing of these fish was estimated to be about £177,000, or less than 1% of the gross value of the fleet's landings.

That cost is based on the disposal of the unmarketable fish to landfill. It should be noted that although this disposal option is available in Shetland it probably will not be an option elsewhere. Most landfill sites in the UK are not licensed to accept fish waste (the site in Shetland is so licensed). Furthermore, from 2021 the disposal of biodegradable waste (including fish) in landfill will be prohibited under the Waste (Scotland) Regulations 2012.

Processing for fish meal represents a possible alternative disposal route for unmarketable discards but this is likely to depend on the quantities and characteristics of the material involved and so probably cannot be depended on.

Substantial questions remain to be answered about how unmarketable fish landed under the discards ban are to be disposed of and where the responsibility for this will lie. Consideration may also need to be given to the disposal of marketable fish that cannot be sold (due to lack of market demand).

5.3.2 Quota Uptake

If fishermen are compelled to stop fishing when any one of their quotas has run out this would prevent them from catching their full quotas of other species in the mixed whitefish fishery and could result in substantial loss of income.

While this analysis suggests that the cod quota would not run out until about October, that would still leave almost one quarter of the Shetland whitefish fleet's total quotas of all species uncaught. Hake and saithe quotas would likely run out much earlier, (July / August) resulting in the loss of about half of the fleet's potential catch of all species. Losses of income on this scale would have a significant impact on the financial viability of the fishing vessels involved. The disruption to fishing patterns implied would also have substantial impacts on fish markets and other industry infrastructure.

Fishermen (and Producer Organisations) can increase the available quota by buying, swapping, leasing or otherwise transferring it, and the figures presented here for 'total available quota' include in-year transfers. In 2014 the Shetland Fish Producer's Organisation and its member vessels transferred in some 1,000 tonnes of additional quota for cod, hake and saithe; accounting for one quarter of the final total available quota for these species (Table 14). For hake, in-year transfers accounted for almost two-thirds of the available quota. However, the availability of quota for such transfers is limited and in-year transfers were not sufficient to cover all of the fish caught. The industry view is that there is simply not enough quota in the system to meet the needs of all fishermen.

This problem with quota availability, and the difficulties of avoiding catches of species for which quota is limited, were highlighted by a trial carried out by Marine Scotland in 2013 (Marine Scotland, 2013) during which a pair of Scottish whitefish vessels attempted to fish under full discard ban conditions. Despite the vessels receiving additional allocations of quota from Marine Scotland the trial had to be terminated early because the skippers were unable either to obtain sufficient quota for the fish they were catching or to avoid catching species for which quota was limited.

Table 14: The basic allocation of quota for selected species to SFPO member vessels in 2014, total in-year transfers and the final total available quota for the year. The in-year changes are shown as percentages of the final available totals. (Data from the SFPO).

Species	Basic Allocation (tonnes)	In-year Change (tonnes)	Total Quota Available (tonnes)	Change as % of Final
Cod	1,794	+705	2,499	+28%
Haddock	4,796	-399	4,397	-9%
Hake	111	+192	304	+63%
Saithe	1,000	+110	1,109	+10%
Whiting	1,885	+243	2,128	+11%
Cod, Hake & Saithe	2,905	+1,007	3,912	+26%