应用贝叶斯方法对有效信息较少的渔业资源进行资源评估-以印度洋长鳍金枪鱼的资源评估为例

众所周知,对有效信息较少的渔业资源进行资源评估面临很大的挑战,而贝叶斯方法在数据数量较少、质量较差的情况下能利用其它种群高质量的数据或已知的先验信息提高资源评估结果的可靠性。由于印度洋长鳍金枪鱼的数据质量较差而数据量有限,长鳍金枪鱼的资源评估结果存在很大的不确定性,为此,本文以印度洋长鳍金枪鱼的资源评估为例,以调查贝叶斯方法在有效信息较少的资源评估中的优势。本文根据不同的先验假设与捕捞数据系列,共构建了8个贝叶斯动态产量模型,以评估长鳍金枪鱼资源。结果表明:(1)分析参数的后验分布能提高捕捞数据系列选择与参数假设的合理性; (2) 利用种群统计学方法为内禀增长率(r)构建有信息先验,能提高资源评估结果的可靠性。与传统方法相比,当基于贝叶斯框架时,能将已知的知识表示为先验信息并能分析参数的后验分布,从而在数据较少或数据质量较差的情况下,能利用各种信息提高参数估计的合理性与资源评估的可靠性。因此,对数据量较少或数据质量较差情况下的渔业资源评估而言,贝叶斯方法非常有效,如本文所示的印度洋长鳍金枪鱼的资源评估。

Application of a Bayesian method to data-poor stock assessment by using Indian Ocean albacore (Thunnus alalunga) stock assessment as an example

It is widely recognized that assessments of the status of data-poor fish stocks are challenging and that Bayesian analysis is one of the methods which can be used to improve the reliability of stock assessments in data-poor situations through borrowing strength from prior information deduced from species with good-quality data or other known information. Because there is considerable uncertainty remaining in the stock assessment of albacore tuna (Thunnus alalunga) in the Indian Ocean due to the limited and low-quality data, we investigate the advantages of a Bayesian method in data-poor stock assessment by using Indian Ocean albacore stock assessment as an example. Eight Bayesian biomass dynamics models with different prior assumptions and catch data series were developed to assess the stock. The results show (1) the rationality of choice of catch data series and assumption of parameters could be enhanced by analyzing the posterior distribution of the parameters; (2) the reliability of the stock assessment could be improved by using demographic methods to construct a prior for the intrinsic rate of increase (r). Because we can make use of more information to improve the rationality of parameter estimation and the reliability of the stock assessment compared with traditional statistical methods by incorporating any available knowledge into the informative priors and analyzing the posterior distribution based on Bayesian framework in data-poor situations, we suggest that the Bayesian method should be an alternative method to be applied in data-poor species stock assessment, such as Indian Ocean albacore.