Inter-benchmark Protocol on Reference Points for Western Horse Mackerel (Trachurus trachurus) in Subarea 8 and Divisions 2.a, 4.a, 5.b, 6.a, 7.a-c,e-k (the northeast Atlantic) (IBPWHM)

The focus of this process was to review and update, if appropriate the stock MSY and PA referencepoints. In addition, alternative assessment models were considered, with emphasis on reducingthe retrospective bias characteristic of the current assessment.The current reference points for Western Horse Mackerel were determined during the 2017 assessmentbenchmark exercise during which an implementation of the SS3 assessment model wasadopted. Based upon the output of the new assessment, stochastic simulations were conductedto derive MSY reference points. The precautionary biomass limit reference point was set to thelowest observed biomass, given no clear indication of impaired recruitment at any observed biomassabove this.Subsequent update assessments in 2017 and 2018 led to revision in the absolute value and timingof the assessment Bloss. Deriving annual catch advice from the most recent assessment output andreference points determined at the benchmark gave rise to annual changes in catch advice atodds with the most recent information on stock development. It is therefore considered that thereference points as estimated in 2017 are no longer suitable and require revision.An extensive range of stochastic simulations using the ICES EqSim software were conducted toexplore a number of scenarios with regard to the stock-recruit relationship, the basis for precautionarybiomass limits, the time period for the underlying source data and the associated assessmentmodel run. The consideration of each of the three SS3 model runs (the benchmark run andsubsequent update assessments) allows the robustness of the reference point values potentialfuture assessment retrospective bias to be determined.As with previous exercises exploring the parameterisation of the stock and recruit relationshipfor the purposes of simulation, no clear functional relationship is evident. A mixed model approachis also inappropriate as the proportions of individual models are variable when individualassessments, time periods and the influence of individual data points are explored. The IBPconcluded that the most appropriate formulation is a segmented regression. However, explorationof fits of this model to the underlying stock and recruit datasets indicated that the breakpointis poorly defined and is particularly sensitive to individual data points. It was considered that asegmented regression, with a breakpoint constrained at the associated Blim value is most appropriatefor the parameterisation of recruitment in a long term simulation of this stock. This conclusionis in agreement with that from the previous benchmark study.The stock assessment is based on data available from 1982. The recruitment estimate from thefirst year is over 20 times the mean of the time series and is 3 times the size of the next largest (in2001). Western horse mackerel is relatively long lived and the 1982 year-class dominated boththe stock and fishery for more than a decade. There is no indication of a likely driver for thisextraordinary year class (including stock size) and no comparable year class has emerged since.The analyses considered 3 different time periods; all years, excluding 1982 and from 1995 on with1995 selected as a cut-off as it corresponds to the year when the contribution of the 1982 yearclassto the spawning stock biomass falls below that of the maximum contribution from the nextlargest year class (2001). During the period when the 1982 year-class was prevalent, SSB rosesignificantly before returning to levels similar to that in 1982. The IBP selected the time periodfrom 1995 on as a basis for the final reference point estimates as it is considered to be a moreappropriate basis for modelling of future recruitment.