Impact Of CPUE Index Weighting On Stock Assessment Using Simulation Study:An Example Of Indian Ocean Albacore Thunnus Alalunga

Albacore tuna（Thunnus alalunga）,as one of the important economic species in tuna fishery,has generally become the research object of experts and scholars in related fields.Its biological characteristics and resource change pattern have been included in the key research by regional international fisheries organizations.The working party on temperate tunas（WPTm T）of the Indian Ocean Tuna Commission（IOTC）has held many meetings to evaluate the resource status of the Indian Ocean albacore tuna using various stock assessment models.However,it is difficult to specify models accurately since all models are simplified representations of real-world situations.Considering the prevalence of data conflicts in the model,the relative weighting of different data sets becomes an essential work.The catch per unit effort（CPUE）index,as an indicator of the abundance of fishery stock,is likewise used as a key factor for data weighting.Therefore,this study was based on the catch,catch-at-age composition and CPUE of albacore tuna in the Indian Ocean from 1980 to 2017,and used the Statistical-Catch-At-Age model ASAP to set up a base model using the corresponding biological parameters to assess the status of albacore tuna in the Indian Ocean.Simulation has been used widely to evaluate the performances of assessment models;therefore,it was used to evaluate the impact of CPUE index weighting on stock assessments.An operating model（OM）was developed to mimic the population dynamics and fishery operating,and estimation models（EMs）conducted by the Statistical-Catch-At-Age model were used to compare the consequence（e.g.,the ratio of depletion）under different CPUE weighting scenarios.Besides,the sensitivity analysis was generated to assess population dynamics,but with mis-specific natural mortality（M）and steepness（h）in the EMs.This study illustrated the results of the weighting selection for the abundance index.Besides,it showed the sensitivity analysis should be conducted to cover possible uncertainty results from imprecise CPUE when specific biological parameters were mis-specified.The results of the study show that the current stock status of the Indian Ocean albacore is not overfished and subject to overfishing:SSB₂₀₁₇/SSB_MSY>1 and F₂₀₁₇/F_MSY>1,which is also the same as the stock assessment result of Indian Ocean albacore by WPTm T in 2019.Sensitivity analysis showed that as the natural mortality（M）increased,the fishing mortality（F）decreased and the spawning parental biomass（SSB）increased.In addition,the results of all models showed that F was generally increasing and SSB was decreasing for Indian Ocean albacore since 1980,perhaps due to the gradual increase in fishing effort for this species since 1980.The results of simulation test show that“high uncertainty in catch”and“low proportion in catch”had a more accurate estimation on F and SSB than the others when M and h could be correctly estimated or even underestimated in the study.Furthermore,the“high uncertainty in CPUE”unexpectedly produced the lowest relative RMSEs for both F and SSB than the other scenarios when M was over specified in EMs.Normally,over-specification M would lead to the expected positive bias for SSB.However,only“high uncertainty in CPUE series”performed differently from the other scenarios,resulting in a negative bias for SSB when M with the correct specification.Therefore,the unexpected performance may be a negative bias from“high uncertainty in CPUE series”,which was partly offset the positive bias from over-specification M.For EMs with over-specification M,“high uncertainty in CPUE series”produced the remarkable lowest MREs for SSB_last/SSB_start;on the contrary,for the other cases,“high uncertainty in CPUE series”resulted in the highest MAREs for SSB_last/SSB_startamong other scenarios.Therefore,fleets or surveys with high uncertainty or not representative catch（low proportion for whole area harvest）perhaps should be assigned higher weighting than fleets with the other drawback（“high uncertainty in CPUE”and“short time series in CPUE”）,especially for supplement could provide confident and informative parameters for assessment.Meanwhile,the confidence for the correct setting of parameters（e.g.,M and h）should be taken into account in CPUE weighting;at least the sensitivity analysis should be conducted to cover the potential model or parameter misspecification and its associated CPUE weighting.However,considering that there is still no objective criterion for the methods of data weighting,the weighting factors of different CPUE data sets are inevitably influenced by subjective factors.Nevertheless,the impact of data weighting on stock assessment is critical and requires careful consideration in model development.