种群模拟在渔业资源评估中的研究现状及展望

随着渔业资源评估理论、数理统计方法和计算机技术的进步, 资源评估模型朝着多样化和复杂化不断发展, 其中种群模拟技术是检测模型适用性和局限性的重要手段。该技术由种群仿真理念发展而来, 通过模拟“真实”种群的方式, 对资源评估结果和管理策略进行有效的评价和预测, 并凭借可结合海洋环境因子、鱼类洄游空间分布以及多鱼种渔业进行资源评估的特性, 已成为开发新资源的重要评估方法之一。为此, 本文对种群模拟的结构和发展过程进行了回顾, 对该技术的核心组成部分操作模型和常见的四类误差(过程误差、观测误差、模型结构误差和管理误差)展开分类讨论。此外, 本文还结合近年来迅速发展的数据缺乏和数据适中模型的特点, 根据实际应用案例对种群模拟的作用和使用前景进行梳理, 并就种群模拟技术发展中存在的主要问题和潜在解决办法提出分析和建议。

Research process and prospects of population simulations in fishery stock assessment

Stock assessment is the key to the scientific management and decision-making of fishery resources. With improvements in stock assessment and computer performance, stock assessments are developing toward diversification and complexity. In the last decade, more than 16 new data-limited/poor stock assessment methods and approaches have been developed for species and fisheries with low-quality data. However, there is a lack of effective tools for developing, selecting, and diagnosing assessment models. Population simulations can promote effective management strategy evaluations by building operating models. Meanwhile, population simulation is an important tool to develop new assessment approaches, due to its ability to flexibly integrate different sources of information, such as environment survey data, tagging and releasing data, and multi-species data. In general, population simulation usually consists of 4 components: the operating models, observation model OM, estimation model, and performance metrics. This study first summarized and reviewed the structure and development of population simulation, then classified and evaluated the advantages and disadvantages of various OMs, namely, a population-dynamics/ surplus-production-model-based OM, statistical catch-at-age models, age structured-stock-assessmentprogram-based OM, and integrated-analysis-based OM. In addition, for the applications of the population simulation, we summarized the potential problems of four common sources of uncertainty in stock assessment and fishery management: (1) process error, indicating the deviation or bias caused by the species itself and the environment; (2) observation error, reflecting the uncertainty caused by fishery operations, data collection, and time-varying catch efficiency; (3) model error, representing the error of parameter estimation and model structure (for example, the equilibrium production model ignores the age structure); and (4) management implementation error, indicating errors generated during the implementation of management measures, such as discard rate, misreporting, and illegal fishing. In addition, we used practical cases to analyze the role and prospect of population simulation in data-limited and data-moderated methods. Simulation analysis can play an effective role in diagnostics when selecting methods on which to base stock assessments. It is suggested that the following issues should be considered and addressed in future research. (1) Most data-limited methods contain potential and essential assumptions, but may not reflect the biological character of the assessed species, which should be considered by decision-makers in model-based management. (2) Multi-species and ecosystem-based population simulation are still under development due to lack of sufficient data, but it will be an important research direction in the future. (3) Neither management strategy evaluation nor stock assessment can completely reject overfishing, and it is possible to select appropriate management strategies with current information and management objectives. Finally, (4) with cross-disciplinary cooperation, population simulation can effectively combine economic, social, conservation, and other factors to formulate more scientific and effective management strategies.