应用贝叶斯状态空间剩余产量模型框架评估印度洋大眼金枪鱼的资源状况

根据1950―2016年的渔获量数据及1955―2016年的单位捕捞努力量(Catch Per Unit Effort,CPUE)数据,采用贝叶斯状态空间剩余产量模型框架JABBA(Just Another Bayesian Biomass Assessment)对印度洋大眼金枪鱼(Thunnus obesus)的资源状况进行评估,分析了渔船效应、CPUE数据尺度对评估结果的影响。结果表明,模型拟合效果对于不同时间跨度下CPUE数据的选择比较敏感。当选用时间跨度为1979―2016年的CPUE数据且考虑渔船效应时,模型拟合效果最好。2016年大眼金枪鱼的资源量为812 kt,最大可持续产量(Maximum Sustainable Yield,MSY)为163 kt,远高于同年渔获量86.81 kt,其资源量具有82.50%的概率处于"健康"状态。当总允许可捕量为69.45～104.17 kt时(2016年渔获量的80%～120%),未来10年大眼金枪鱼的资源量仍高于B_(MSY)(达到MSY所需的生物量)。回顾性分析结果表明,该资源评估结果存在一定程度的回顾性问题,捕捞死亡率和资源量分别存在被低估和高估的现象。将来需要在模型结构设定、CPUE数据选择及模型参数的先验分布设置等方面进一步优化。

Stock assessment for bigeye tuna (Thunnus obesus) in the Indian Ocean using JABBA

Bigeye tuna () is one of the most valuable tropical tuna species targeted by most longline fisheries. stock assessments have always been the focus for regional tuna fishery management organizations worldwide. Based on the catch from 1950 to 2016 and Catch Per Unit Effort (CPUE) from 1955 to 2016, the stock of the Indian Ocean bigeye tuna was assessed by the Bayesian state space surplus production model in an open environment, JABBA (Just Another Bayesian Biomass Assessment), and the implications on the effects of fishing boat and CPUE data scale was explored. The results showed that the stock assessment was sensitive to different CPUE, and the scenario using CPUE considering vessel effect from 1979 to 2016 was revealed to perform best with the lowest Root-Mean-Squared-Error (RMSE) and Deviance Information Criterion (DIC), and selected to be the base case. The median estimate for bigeye tuna biomass in 2016 was 812 kt, and the Maximum Sustainable Yield (MSY) was estimated to be 163 kt, which was much higher than the catch (86.81 kt) in 2016, indicating that the stock was not overfished, with 81% in the green zone of the Kobe plot. The biomass of bigeye tuna would be higher than the biomass that can produce the maximum sustainable yield () in the next 10-year projection when the total permissible catch was set to 69.45-104.17 kt (80%-120% of catch in 2016). There were some retrospective errors in the stock assessment results, with underestimated fishing rate and overestimated biomass. Therefore, the stock assessments should be improved by updating the model structure, CPUE standardization, and setting for prior distribution of model parameters.