气候变化对东南太平洋智利竹筴鱼渔获量的影响

为了探讨气候变化对智利竹筴鱼(Trachurus murphyi)渔获量的长期影响, 采集 1900—2016 年北大西洋涛动

(North Atlantic Oscillation, NAO)、太平洋年代际涛动(Pacific Decadal Oscillation, PDO)、厄尔尼诺(El Ni?o)等 8 个

低频气候变化参数, 全球海气温度异常指标时间序列数据和 1970—2016 年东南太平洋智利竹筴鱼总渔获量数据,

在对其进行相关性分析的基础上, 运用 BP 神经网络模型构建了东南太平洋智利竹筴鱼渔获量预测模型, 并以效率

系数为评价规则对预测模型进行评价, 进而得到了最优预测模型。最后对最优预测模型进行了因子敏感性分析, 提

取了对东南太平洋智利竹筴鱼(Trachurus murphyi)影响较大的因子。最优预测模型拟合效果显示, 渔获量拟合值与

观测值有基本一致的变化趋势, 两个序列的线性相关系数为 0.745, 模型拟合效果良好。最优模型因子敏感性分析

表明, 在研究期间, 影响东南太平洋智利竹筴鱼渔获量的气候变化表征因子主要为北大西洋涛动、太平洋年代际涛

动和北太平洋指数。

Impact of climate change on Chilean jack mackerel catch in the Southeast Pacific

Chilean jack mackerel (Trachurus murphyi) is a species of pelagic fish widely distributed in the subtropical waters of the South Pacific Ocean. It is one of the main commercial fish species in the Southeast Pacific Ocean. The catch of Chilean jack mackerel began to increase steadily in the 1970s, decreased rapidly after reaching its peak in 1995, and then remained at a low level. This may be due to many reasons, including global climate change; thus, thepotential climatic reasons for the change in Chilean jack mackerel catch will be discussed in this paper. The long-time series data of eight low-frequency climate change parameters, such as the North Atlantic Oscillation (NAO), Pacific Decadal Oscillation (PDO), North Pacific Index (NPI )and El Ni?o, from 1900 to 2016; and the global sea air temperature anomaly index and total catch data for Chilean jack mackerel in the Southeast Pacific from 1970 to 2016 were collected. Based on correlation analysis of the above data, prediction models of Chilean jack mackerel catch in the Southeast Pacific Ocean were established using a back propagation (BP) neural network model, and by taking the efficiency coefficient as the evaluation rule, the optimal prediction model was obtained. The results showed that the optimal prediction model was a three-layer BP neural network with six neurons in the hidden layer; in the optimal model, the fitting and observed value of catch had basically the same trend; and the linear correlation coefficient of the two sequences was 0.745. The fitting effect of the optimal prediction model was good. The factor sensitivity analysis of the optimal prediction model showed that during the study period, the NAO, PDO, andNPI are the main factors that affect the catch of Chilean jack mackerel in the Southeast Pacific Ocean. In the existing research, on the one hand, the factors affecting Chilean jack mackerel fisheries are usually the elements of the marine environment affected by climate change; there are few discussions on regional or global climate variables. On the other hand, most current studies are short-term studies within 10 years; however, climate change often lasts for long periods of time (usually for decades or longer). Therefore, these studies have limitations in revealing the impact of climate change on fisheries. In this study, based on the long-term (more than 100 years) data of low-frequency climate change parameters, a prediction model of Chilean jack mackerel catch in the Southeast Pacific Ocean was constructed using a BP neural network model. It can help to analyze the long-term impact of climate change on Chilean jack mackerel fishery resources from the perspective of global climate change, and provide scientific basis for the sustainable development of Chilean jack mackerel fishery.