Stock Assessment And Management Strategy Evaluation Of Striped Marlin (Tetrapturus Audax) In The Indian Ocean

Billfishes are typically not the targets of large-scale fisheries,apart from swordfish ?Xiphias gladius?,which historically lead to a lack of management procedures on these species.Striped marlin?Tetrapturus audax?is known as lacking age-structured data and abundance index data since it is considered a by-catch species.It also suffers from difficulties of identification,highly biased catch data and environmental impacts on stock dynamics.This hard situation brings a lot of challenges in assessing and managing the species.In this study,we organized and analyzed assessment methods of billfish species which were used in workshops and working parties of Regional Fisheries Management Organizations?RFMOs?.We discussed the progress of stock assessment and achievements of billfish in terms of data availability,assessment method,model structure and hypothesis,respectively.We also made comparisons on different models used by different researchers through years.Then we connected Management Strategy Evaluation?MSE?with stock assessment to build a bridge between assessment and management.There are basically three main sections of an MSE,operating model,observation model and implementation model.MSE structure is flexible with any change due to the real condition of a fishery or stock status and it does good job in quantifying fisheries management.We conducted stock assessment and MSE of striped marlin in Indian Ocean in our study.As a result,stock status was assessed while related parameters and reference points were calculated.The study was conducted with time-series data while making trade-offs on management procedures to provide management advice on the fishery.The best available management procedures are chosen under quantitative analysis and comes with Total Allowable Catch?TAC?as a result.Afterwards,we started Value of Information?VOI?analysis based on observation model.The performance and efforts made by data were detected using quantitative analysis and advices and suggestions were made based on the results.We chose delay difference model to assess striped marlin stock in the Indian Ocean based on the analysis we did in review combined with consideration of data availability and stock and fisheries status.Slight modifications were made to achieve better expression of stock dynamics in this state-space age-structured delay difference model.Specially,we insert an incidence function in the assessment model and natural mortality,age-structured length and weight,maturity and vulnerability are defined in the function.Five standard CPUE time series were used to build five assessment models,named JPSW,JPNE,JPNW,TWNE and TWNW.We conduct the calculation using solver function in Excel based on maximum likelihood estimation.In the result of model fit of the five models,B₀ ranges with 95%confidence interval are,27.98?17.0,56.4?,28.09?17.0,30.6?,28.67?27.0,32.6?,17.15?0,19.8?and 38.34?32.0,56.4?,and steepness values with 95%confidence interval are,0.83?0.71,0.97?,0.71?0.66,0.78?,0.8?0.78,0.82?,0.71?0.68,0.74?and 0.53?0.35,0.68?.We also discussed correlations of Maximum Sustainable Yield?MSY?,harvest rate U_MSY and stock dynamics.MSY results of the models we built are4.16,3.63,4.12,2.21 and 3.73 while U_MSY values are 0.45,0.36,0.43,0.36 and 0.24.We also calculated B_MSY with the results 10.19,11.16,10.70,6.82 and 16.96 came out.Sensitivity analysis were conducted on parameters including type of stock recruitment relationship,natural mortality,maturity and vulnerability curve shape.Then we discussed the robustness of our model and weakness of our assessment work.Based on our results in the delay difference assessment model combined with stock assessment result of Indian Ocean Tuna Commission?IOTC?Working Party on Billfish?WPB16?,we built up a state-space age-structured operating model in which stock dynamics of striped marlin in Indian Ocean was expressed.Twenty one Management Procedures?MPs?were chosen to be tested in our MSE simulation circle with 48simulations for each MP and 50-year projection of yield.The result of convergence analysis told us yield results are constantly stable within 48 simulations.Trade-offs were made with Kobe plots and eventually 11 MPs were chosen for further analysis and they are,CC1,DD,SBT1,GB?slope,DD4010,cur E,MRreal,minlen Lopt1,Etarget Lopt,matlenlim and MRnoreal,in which we have 5 output-control MPs.Consequently,we found DD4010 as the best practise MP throughout reference point adjustment and choices of different harvest control rules.Then we calculated TAC using two standard abundance index time-series from Japan and Taiwan,China.TAC results were calculated under the implementation of 5 output-control MPs,DD4010,GB?slope,SBT1,CC1 and DD.TAC results under the implementation of the best practice MP,DD4010 are,2694?1554.47,3415.38?and 2802.77?1662.70,4115.08?with 50%confidence interval,respectively.Last step,we conducted sensitivity analysis on every parameter by MPs.Model is apparently sensitive to abundance index and natural mortality related parameters with implementations of different MPs.We conducted value of information analysis on annual catch error,annual index error,catch bias,natural mortality bias,von Bertalanffy growth parameter K bias,maximum length bias and catch-at-age sample size.The analysis is carried out with redefining distribution of parameter errors or biases through observation model in MSE.We gave the range of?0.05,0.15?to annual catch error and annual index error while the range of catch bias and maximum length bias is?4/5,5/4?and the range of natural mortality and von Bertalanffy growth parameter is?2/3,3/2?.At the same time,we provide a range of?10,1000?to catch-at-age sample size in our simulation.All parameters are assumed a uniform distribution in the defined range.A series of statistical catch-at-age data-rich management procedures are applied in our simulation system.Nine iterations are applied in each parameter range with 128 simulations and 50-year projection of yield while the average of last ten years were used as final yield.Among the seven tested parameters,annual catch error,annual index error and catch bias show relatively low information value with slight effort on yield results.Whereas natural mortality bias,von Bertalanffy growth parameter K and maximum length then show somehow high value of information through significant impact on the results in terms of their precision and quality.The last parameter,catch-at-age sample size,then appeared changing the result at the beginning of its defined range,hence we concluded appropriate sample size of 124 for catch-at-age.Consequently,we started from stock assessment with state-space age-structured delay difference model based one IOTC catch time series and five time-series standard abundance index from Japan and Taiwan,China.The stock we identified overfished and is suffering overfishing as well.We urgently need actions to be taken to prevent overfished and overfishing due to the stock status right now and a stock rebuild plan will also be helpful.We detected DD4010 as the best practice management procedure among alternative ones through trade-offs in MSE simulation test system.TACs were then calculated with management advices.We also analyzed the value of information of some certain parameters via MSE simulation test using DLMtool and MSEtool,in which how information contribute to the results was quantified and we make advices and conclusions on scientific resources and effort arrangement.In general,we built our model based on the fishery of striped marlin in the Indian Ocean,then conducting the study with simulation test,statistical analysis to assess the target stock and evaluate the performance of alternative management procedures.Practices were made for detecting stock status with fisheries conservation and management advices based on the best available science.This work will also benefit to achieve the same target for fisheries managers and stakeholders.And this study provides valuable information specific on future data collection and opens a new window of reconstruction of scientific resources and efforts.