根据2009—2012年南太平洋长鳍金枪鱼（Thunnus alalunga）延绳钓生产统计数据及遥感获取的海表温度（sea surface temperature,SST）、叶绿素a浓度（chlorophyll a concentration,Chl-a）和海面高度距平（sea surface height anomaly,S...根据2009—2012年南太平洋长鳍金枪鱼（Thunnus alalunga）延绳钓生产统计数据及遥感获取的海表温度（sea surface temperature,SST）、叶绿素a浓度（chlorophyll a concentration,Chl-a）和海面高度距平（sea surface height anomaly,SSHA）等环境数据,分析了长鳍金枪鱼单位捕捞努力量渔获量（catch per unit of fishing effort,CPUE）的时空分布及其与环境因子的相关性。结果表明：长鳍金枪鱼作业渔场主要集中在4°S—28°S、158°E—176°E附近海域;长鳍金枪鱼渔场CPUE呈明显的季节性变化,1—3月CPUE值较低（〈12.5尾·千钩-1）,随后逐渐增加,至7月达到最大值为18.1尾·千钩-1,而8—12月基本呈逐渐降低趋势;1月渔场重心位于16°S、168°E附近海域,2—3月向西北偏移,而在3—7月逐渐向东南方向转移,8月以后开始逐渐回撤至西北方向,在9—12月渔场重心变化幅度相对较小,主要位于15°S—16°S、168°E—169°E海域;总体来说,长鳍金枪鱼中心渔场最适SST为27.0～30.5℃,次适SST为20～24℃;最适叶绿素a浓度为0.02～0.08mg·m-3,最适海面高度距平为3～23 cm。展开更多
南太平洋长鳍金枪鱼是我国远洋渔业的重点捕捞对象,对南太平洋长鳍金枪鱼进行准确的渔场预报,可以提高捕捞效率,提高渔业的生产能力。本研究根据1993-2010年南太平洋长鳍金枪鱼的延绳钓生产数据以及海洋卫星遥感数据(海水表面温度,SST;...南太平洋长鳍金枪鱼是我国远洋渔业的重点捕捞对象,对南太平洋长鳍金枪鱼进行准确的渔场预报,可以提高捕捞效率,提高渔业的生产能力。本研究根据1993-2010年南太平洋长鳍金枪鱼的延绳钓生产数据以及海洋卫星遥感数据(海水表面温度,SST;海面高度,SSH)和ENSO(ElNinoSouthern Oscillation)指标,采用DPS(data processing system)数据处理系统中的BP人工神经网络模型,以渔获产量(单位时间的渔获尾数)和单位捕捞努力量渔获量(CPUE,Catch per unit of effort)分别作为中心渔场的表征因子,并作为BP模型的输出因子,以月、经度、纬度、SST、SSH和ENSO指标等作为输入因子,分别构建4-3-1,5-4-1,5-3-1,6-5-1,6-4-1,6-3-1等BP模型结构,比较渔场预报模型优劣。研究结果表明,以CPUE作为输出因子的BP人工神经网络结构总体上较优,其中以6-4-1模型结构为最优,相对误差只有0.006 41。研究认为,以CPUE为输出因子的6-4-1结构的人工神经网络模型,能够准确预报南太平洋长鳍金枪鱼的渔场位置。展开更多
为提高渔场资源丰度预测准确性,以单位捕捞努力量渔获量(catch per unit effort,CPUE)为长鳍金枪鱼(Thunnus alalunga)资源丰度指标,利用海洋遥感、Argo等获取的海洋环境因子,在最优分布式决策梯度提升树(XGBoost)模型基础上,采用卷积...为提高渔场资源丰度预测准确性,以单位捕捞努力量渔获量(catch per unit effort,CPUE)为长鳍金枪鱼(Thunnus alalunga)资源丰度指标,利用海洋遥感、Argo等获取的海洋环境因子,在最优分布式决策梯度提升树(XGBoost)模型基础上,采用卷积神经网络(Convolutional Neutral Network,CNN)进行高维海洋环境数据特征提取,并利用模拟退火算法(Simulate Anneal,SA)对最优分布式决策梯度提升树(Extreme gradient boosting,XGBoost)模型进行超参数优化,提出了改进的XGBoost模型CNN-SA-XGBoost模型,实现对南太平洋长鳍金枪鱼资源丰度的回归预测。实验表明,在南太平洋长鳍金枪鱼资源丰度预测中,CNN-SA-XGBoost模型的均方根误差为0.486,较XGBoost减少12.4%,较多元线性回归(Multiple Linear Regression)、随机森林(Random Forest,RF)、BP神经网络等模型预测误差降低11.8%~28.4%。且改进的XGBoost模型在一定程度上改善了传统资源丰度预测模型面对高维环境数据和缺失值较多的渔业生产数据时预测误差较大的问题,为远洋渔场预报提供了新方法。展开更多
Over the years there has been growing interest regarding the effects of climatic variations on marine biodiversity. The exclusive economic zones of South Pacific Islands and territories are home to major international...Over the years there has been growing interest regarding the effects of climatic variations on marine biodiversity. The exclusive economic zones of South Pacific Islands and territories are home to major international exploitable stocks of albacore tuna (Thunnus alalunga);however the impact of climatic variations on these stocks is not fully understood. This study was aimed at determining the climatic variables which have impact on the time series stock fluctuation pattern of albacore tuna stock in the Eastern and Western South Pacific Ocean which was divided into three zones. The relationship of the climatic variables for the global mean land and ocean temperature index (LOTI), the Pacific warm pool index (PWI) and the Pacific decadal oscillation (PDO) was investigated against the albacore tuna catch per unit effort (CPUE) time series in Zone 1, Zone 2 and Zone 3 of the South Pacific Ocean from 1957 to 2008. From the results it was observed that LOTI, PWI and PDO at different lag periods exhibited significant correlation with albacore tuna CPUE for all three areas. LOTI, PWI and PDO were used as independent variables to develop suitable stock reproduction models for the trajectory of albacore tuna CPUE in Zone 1, Zone 2 and Zone 3. Model selection was based on Akaike Information Criterion (AIC), R2 values and significant parameter estimates at p < 0.05. The final models for albacore tuna CPUE in all three zones incorporated all three independent variables of LOTI, PWI and PDO. From the findings it can be said that the climatic conditions of LOTI, PWI and PDO play significant roles in structuring the stock dynamics of the albacore tuna in the Eastern and Western South Pacific Ocean. It is imperative to take these factors into account when making management decisions for albacore tuna in these areas.展开更多
文摘根据2009—2012年南太平洋长鳍金枪鱼（Thunnus alalunga）延绳钓生产统计数据及遥感获取的海表温度（sea surface temperature,SST）、叶绿素a浓度（chlorophyll a concentration,Chl-a）和海面高度距平（sea surface height anomaly,SSHA）等环境数据,分析了长鳍金枪鱼单位捕捞努力量渔获量（catch per unit of fishing effort,CPUE）的时空分布及其与环境因子的相关性。结果表明：长鳍金枪鱼作业渔场主要集中在4°S—28°S、158°E—176°E附近海域;长鳍金枪鱼渔场CPUE呈明显的季节性变化,1—3月CPUE值较低（〈12.5尾·千钩-1）,随后逐渐增加,至7月达到最大值为18.1尾·千钩-1,而8—12月基本呈逐渐降低趋势;1月渔场重心位于16°S、168°E附近海域,2—3月向西北偏移,而在3—7月逐渐向东南方向转移,8月以后开始逐渐回撤至西北方向,在9—12月渔场重心变化幅度相对较小,主要位于15°S—16°S、168°E—169°E海域;总体来说,长鳍金枪鱼中心渔场最适SST为27.0～30.5℃,次适SST为20～24℃;最适叶绿素a浓度为0.02～0.08mg·m-3,最适海面高度距平为3～23 cm。
文摘南太平洋长鳍金枪鱼是我国远洋渔业的重点捕捞对象,对南太平洋长鳍金枪鱼进行准确的渔场预报,可以提高捕捞效率,提高渔业的生产能力。本研究根据1993-2010年南太平洋长鳍金枪鱼的延绳钓生产数据以及海洋卫星遥感数据(海水表面温度,SST;海面高度,SSH)和ENSO(ElNinoSouthern Oscillation)指标,采用DPS(data processing system)数据处理系统中的BP人工神经网络模型,以渔获产量(单位时间的渔获尾数)和单位捕捞努力量渔获量(CPUE,Catch per unit of effort)分别作为中心渔场的表征因子,并作为BP模型的输出因子,以月、经度、纬度、SST、SSH和ENSO指标等作为输入因子,分别构建4-3-1,5-4-1,5-3-1,6-5-1,6-4-1,6-3-1等BP模型结构,比较渔场预报模型优劣。研究结果表明,以CPUE作为输出因子的BP人工神经网络结构总体上较优,其中以6-4-1模型结构为最优,相对误差只有0.006 41。研究认为,以CPUE为输出因子的6-4-1结构的人工神经网络模型,能够准确预报南太平洋长鳍金枪鱼的渔场位置。
文摘为提高渔场资源丰度预测准确性,以单位捕捞努力量渔获量(catch per unit effort,CPUE)为长鳍金枪鱼(Thunnus alalunga)资源丰度指标,利用海洋遥感、Argo等获取的海洋环境因子,在最优分布式决策梯度提升树(XGBoost)模型基础上,采用卷积神经网络(Convolutional Neutral Network,CNN)进行高维海洋环境数据特征提取,并利用模拟退火算法(Simulate Anneal,SA)对最优分布式决策梯度提升树(Extreme gradient boosting,XGBoost)模型进行超参数优化,提出了改进的XGBoost模型CNN-SA-XGBoost模型,实现对南太平洋长鳍金枪鱼资源丰度的回归预测。实验表明,在南太平洋长鳍金枪鱼资源丰度预测中,CNN-SA-XGBoost模型的均方根误差为0.486,较XGBoost减少12.4%,较多元线性回归(Multiple Linear Regression)、随机森林(Random Forest,RF)、BP神经网络等模型预测误差降低11.8%~28.4%。且改进的XGBoost模型在一定程度上改善了传统资源丰度预测模型面对高维环境数据和缺失值较多的渔业生产数据时预测误差较大的问题,为远洋渔场预报提供了新方法。
文摘Over the years there has been growing interest regarding the effects of climatic variations on marine biodiversity. The exclusive economic zones of South Pacific Islands and territories are home to major international exploitable stocks of albacore tuna (Thunnus alalunga);however the impact of climatic variations on these stocks is not fully understood. This study was aimed at determining the climatic variables which have impact on the time series stock fluctuation pattern of albacore tuna stock in the Eastern and Western South Pacific Ocean which was divided into three zones. The relationship of the climatic variables for the global mean land and ocean temperature index (LOTI), the Pacific warm pool index (PWI) and the Pacific decadal oscillation (PDO) was investigated against the albacore tuna catch per unit effort (CPUE) time series in Zone 1, Zone 2 and Zone 3 of the South Pacific Ocean from 1957 to 2008. From the results it was observed that LOTI, PWI and PDO at different lag periods exhibited significant correlation with albacore tuna CPUE for all three areas. LOTI, PWI and PDO were used as independent variables to develop suitable stock reproduction models for the trajectory of albacore tuna CPUE in Zone 1, Zone 2 and Zone 3. Model selection was based on Akaike Information Criterion (AIC), R2 values and significant parameter estimates at p < 0.05. The final models for albacore tuna CPUE in all three zones incorporated all three independent variables of LOTI, PWI and PDO. From the findings it can be said that the climatic conditions of LOTI, PWI and PDO play significant roles in structuring the stock dynamics of the albacore tuna in the Eastern and Western South Pacific Ocean. It is imperative to take these factors into account when making management decisions for albacore tuna in these areas.
