摘要
为探讨气候变化对长鳍金枪鱼渔获量的影响,利用中西太平洋渔业委员会统计的 1960-2021年太平洋长鳍金枪鱼年度渔获量和对应的厄尔尼诺指标(Niño1+2、Niño3、Niño4以及Niño3.4)、南方涛动指数(SOI)、北大西洋涛动(NAO)、太平洋年代际涛动(PDO)、北太平洋指数(NPI)以及全球海气温度异常指标(dT)等月度数据,采用BP神经网络和变量敏感性分析法探讨了低频气候因子与长鳍金枪鱼渔获量的关系;构建了结构为 6-8-1的最优BP神经网络模型,对长鳍金枪鱼渔获量进行了预测.结果表明,Niño1+2、SOI、NAO、PDO、NPI、dT为影响长鳍金枪鱼渔获量相对独立的气候因子,其对应的最佳滞后阶数依次为 8年、2年、9年、0年、9年、3年.Niño1+2、SOI、NAO为影响长鳍金枪鱼渔获量的关键气候因子.长鳍金枪鱼渔获量预测值和实际值差值与实际值的比值自 1971年后基本维持在15%以内,预测值与实际值变化趋势基本一致,模型拟合效果良好.
Abstract
In order to investigate the impact of climate change on the catch of bigeye tuna,we utilized the annual Pacific bigeye tuna catch data from 1960 to 2021,which was statistically compiled by the Western and Central Pa-cific Fisheries Commission.We also employed corresponding monthly climate indices,including Niño1+2,Niño3,Niño4,Niño3.4,Southern Oscillation Index(SOI),North Atlantic Oscillation(NAO),Pacific Decadal Oscillation(PDO),North Pacific Index(NPI),and global sea-air temperature anomaly(dT).By using a BP neural network and variable sensitivity analysis,we examined the relationship between these low-frequency climate factors and bigeye tuna catch.Our findings revealed that Niño1+2,SOI,NAO,PDO,NPI,and dT are relatively independent climate factors that have an impact on bigeye tuna catch.The optimal lag orders for these climate factors were determined to be 8 years for Niño1+2,2 years for SOI,9 years for NAO,0 years for PDO,9 years for NPI,and 3 years for dT.Among these factors,Niño1+2,SOI,and NAO were identified as the key climate factors influencing bigeye tuna catch.We constructed an optimal BP neural network model with a structure of 6-8-1,and the ratio of the difference between the predicted and actual bigeye tuna catch to the actual catch has been maintained within 15%since 1971.Additionally,the trend of the predicted and actual catch was found to be basically consistent,indicating a satisfact-ory level of model fit.
NETL
NSTL
万方数据