Evaluation of a multi-annual plan including an index based HCR for North Sea horse mackerel

For most pelagic fish stocks that are exploited in EU waters, long-term management plans are established and used for TAC setting. North Sea horse mackerel was classi-fied by ICES as a data-limited stock (DLS) in category 5 with only landings data avail-able in 2012. This led to the advice for 2013 being a TAC which represented a reduction of 20% of recent landings as a precautionary measure. In response, the main stake-holder in the fishery - the Dutch Pelagic Freezer-trawler Association (PFA) - sought collaboration with IMARES to develop a multi-annual plan for this stock. The plan should specify a rationale for establishing TACs through a Harvest Control Rules (HCR) and it should stipulate how the stock could be ‘moved up the ladder’ of the DLS categories; ultimately aiming to progress it to become a data rich (category 1) stock in the medium term.
As a DLS, no stock assessment has previously been adopted as a basis for advice for North Sea horse mackerel. In recent years, some explorative work on data availability for the stock has been done by the ICES Working Group on Widely Distributed Stocks (WGWIDE), and this work has been taken further in this report to produce two potential indices of stock size from the IBTS Q3 survey data. A thorough analysis of the IBTS data for North Sea horse mackerel is also presented. 

A modelling framework was developed (the JAXass model) to obtain plausible estimates of stock size using the indices created from the IBTS data. There are numerous difficulties fitting an assessment model for this stock: unclear stock boundaries, difficulty aging horse mackerel, lack of strong cohort signals in the catch-at-age data, no targeted survey of abundance for North Sea horse mackerel, and the available survey data not covering one of the main fishing grounds for the stock (ICES Subarea VIId). In addition there are some conflicting signals between the catch and index data, espe-cially since the mid-2000s. Hence no single ‘best’ assessment model has been provided. Rather a suite of plausible models are used together to test the robustness of the stock to potential future management options.
Fishing mortality reference point values were calculated by running long-term projections at varying levels of F. FMSY (the fishing mortality providing the maximum yield at equilibrium) is estimated to be in the range 0.08-0.24, depending on the choice of stock recruit relationship. A more precautionary and better estimated robust proxy for FMSY is proposed: F35 (the fishing mortality that leads to SSB at 35% of virgin biomass at equilibrium). Values for F35 range from 0.1-0.14 depending on the choice of stock–recruit relationship. 

The available data suggest that the North Sea horse mackerel stock is currently at a low biomass in relation to levels in the late 1990s. Recent IBTS surveys have had a large proportion of 0 hauls and low cpues. Alternative methods of calculating the index from the raw data (e.g. delta-lognormal method, GLM analysis or simple mean cpue of all hauls) indicate that the stock is near the lowest level observed. As a result, all explora-tory assessment models fit to these data estimate that current SSB is near the lowest observed. Fishing mortality has been above fishing mortality levels suggested for Fmsy (proxies), especially in the recent past and a reduction in catches in the short term is therefore required to allow for recovery of the stock.
The special request sent to ICES asked “to assess whether the proposal as presented in Annex A is precautionary and if it will deliver Fmsy by 2015 or 2020 at the latest”. It was decided at the WKHOMMP meeting that with simulations including a constant catch or a constant F, the focus of the evaluations should rather be on what level the catch could be in the short term in order to reverse the declining trend in the stock than on the performance of the HCRs. To this end, the following procedure was followed: 1. Scenarios of constant catch and constant F were examined to see what level of catch or F would allow for recovery of the stock in the short term. 2. On the basis of these runs, viable short term management options are presented. 3. An initial examination of the potential HCR options is also presented, but to allow for a meaningful compari-son of these some stock recovery is required first.
HCR performance statistics were tailored to short-term considerations to give an idea of the short-term prospects for the fishery and the short-term recovery of the stock (by 2020). A recovery of the stock above SSB2012 (lowest observed, candidate for Blim) is considered the first priority. This recovery is unlikely by 2015, but should be expected at least by 2020. A re-evaluation of potential HCRs looking at longer term management considerations should be undertaken in 3-5 years, unless anything changes in our un-derstanding of the stock in the meantime that would require a re-evaluation of man-agement options.
In terms of providing an estimation of what level of catch the stock could take in the short term, the MSE results are highly sensitive to the choice of stock–recruit relationship (SRR). However, regardless of the choice of SRR, the results from the exploratory assessment model runs and the evaluation of expected future performance of the stock are rather pessimistic. Depending on the SSR scenario (segRick or decLim), the suggested level of catch in the next few years which could allow for a reversal in the de-clining trend in SSB and bring SSB to above the SSB2012 level ranges from 2.5kt to 10kt. Recent catches of this stock have been in the range of 20-25kt, so this represents a significant reduction in fishing opportunities. 

Determining what levels of TACs will lead to this level of catch is problematic, because of the consistent underutilization in the TAC observed in recent years due to the ab-sence of the Danish fishery. In order to reduce the fishing opportunities for those fleets who do utilize their quota not more drastically than necessary, this underutilization could be taken into account when establishing the level of the TAC. In addition, a reduction in TAC would naturally only be effective if the TAC would be restrictive to the collective of all countries exploiting the stock, i.e. including Norway, which has caught highly variable but occasionally substantial quantities.
Investigation of how the HCRs performed compared with each other in the long term, was investigated as a hypothetical exercise however. The current evaluation results, suggest that the HCR based on a log-linear regression interpretation of the slope with a lambda value of 0.5 performs relatively well. It showed a lower interannual variabil-ity than when a 5 year slope was used but still responded most quickly to changes in the trend of the index. This is most likely because, at the time when the declining trend in the SSB starts to reverse and in fact is level for a few years, the 3-year slope HCR keeps the TAC also stable, while the 5-year HCR continues longer with declines, after which quickly changes to increases again. At the same time it indicated more recovery in SSB than the DLS rule or the HCR including a 20% constraint in TAC change.  
There is some anecdotal evidence suggesting that one or more relatively strong year-classes may have been produced in recent years in the North Sea: pelagic fishers have reported substantial quantities of juvenile horse mackerel in the Southern North Sea during their fishery, a near shore survey conducted in the ‘Voordelta’ area by IMARES caught a large number of juvenile horse mackerels in the summer of 2013, and the by-catch quota provided to the Dutch demersal fleet was exhausted approximately midyear in 2013, which was considered very early compared with other years. If these are a true reflection of actual events, then this may lead to a more rapid recovery of the stock than the results of the current study suggest. This should then be measurable in the IBTS survey data in upcoming years, in which case management measures can be revised accordingly.  
Generally speaking, any long-term management plan is most effective when its measures apply to all fisheries exploiting the stock and when catches can be identified as originating from that stock with some certainty. Considering the potential of mixing between Western and North Sea horse mackerel occurring in area VIId, better insight in the origin of landings from that area will be a major benefit, if not crucial, for im-provement of the quality of future scientific advice and thus management of the North Sea horse mackerel stock. One way of possibly distinguishing between individuals of the two stocks is with the GCxGC-MS (Gas chromatography x Gas chromatography–mass spectrometry). A pilot project aimed at determining whether this technique could be used for distinguishing between Western and North Sea horse mackerel is currently underway.