文章利用2008—2015年南太平洋长鳍金枪鱼(Thunnus alalunga)延绳钓渔业数据,结合11个环境指标(海表温度、叶绿素a(Chl-a)浓度、海表温度距平、叶绿素距平、海表温度梯度、叶绿素梯度、海平面异常以及渔区格网对应的前后各1个月海表温...文章利用2008—2015年南太平洋长鳍金枪鱼(Thunnus alalunga)延绳钓渔业数据,结合11个环境指标(海表温度、叶绿素a(Chl-a)浓度、海表温度距平、叶绿素距平、海表温度梯度、叶绿素梯度、海平面异常以及渔区格网对应的前后各1个月海表温度和叶绿素值)和3个时空指标(月、经度和纬度),并基于6种集成学习模型,以月为时间分辨率、0.5°×0.5°为空间分辨率,开展了南太平洋长鳍金枪鱼渔场模型构建和预报研究。模型通过10折交叉验证和网格搜索思想确定最佳参数,采用的随机森林、Bagging决策树、C5.0决策树、梯度提升决策树、AdaBoost、Stacking集成模型分别取得了75.52%、73.87%、72.99%、71.14%、71.33%、75.84%的分类准确率。经对比,Stacking集成模型准确率最高。利用2015年环境数据进行预报精度检验,预报总体准确率为63.86%~82.14%,平均70.99%;高单位捕捞努力量渔获量(Catch per unit effort,CPUE)渔区预报准确率为62.71%~97.85%,平均78.76%。结果表明Stacking集成模型对南太平洋长鳍金枪鱼渔场的预报具有较好的效果及可行性。展开更多
The characteristic changes of South Pacific sea surface temperature anomalies （SSTAs） for the period January 1979 to December 2011, during which the 1990s Pacific pan-decadal variability （PDV） interdecadal regime ...The characteristic changes of South Pacific sea surface temperature anomalies （SSTAs） for the period January 1979 to December 2011, during which the 1990s Pacific pan-decadal variability （PDV） interdecadal regime shifts occurred, were examined. Empirical Orthogonal Function （EOF） analysis was applied to the monthly mean SSTA for two sub-periods： January 1979 to December 1994 （P 1） and January 1996 to December 2011 （P2）. Both the spatial and temporal features of the leading EOF mode for P1 and P2 showed a remarkable difference. The spatial structure of the leading EOF changed from a tripolar pattern for P 1 （EOF-P 1） to a dipole-like pattern for P2 （EOF-P2）. Besides, EOF-P 1 （EOF-P2） had significant spectral peaks at 4.6 yr （2.7 yr）. EOF-P2 not only had a closer association with E1 Nifio-Southern Oscillation （ENSO）, but also showed a faster response to ENSO than EOF-P1 based on their lead-lag relationships with ENSO. During the development of ENSO, the South Pacific SSTA associated with ENSO for both PI and P2 showed a significant eastward propagation. However, after the peak of ENSO, EOF-P1 showed a stronger persistence than EOF-P2, which still showed eastward propagation. The variability of the SSTA associated with the whole process of ENSO evolution during P1 and the SSTA associated with the development of ENSO during P2 support the existence of ocean-to-atmosphere forcing, but the SSTA associated with the decay of ENSO shows the phenomenon of atmosphere-to-ocean forcing.展开更多
长鳍金枪鱼(Thunnus alalung)是南太平洋金枪鱼渔业的主要捕捞种类之一,具有很高的经济价值,掌握其渔场分布与主要海洋环境因子关系是高效开发利用的基础。根据2011年我国南太平洋渔业生产数据和遥感环境数据,按周分析了长鳍金枪鱼单位...长鳍金枪鱼(Thunnus alalung)是南太平洋金枪鱼渔业的主要捕捞种类之一,具有很高的经济价值,掌握其渔场分布与主要海洋环境因子关系是高效开发利用的基础。根据2011年我国南太平洋渔业生产数据和遥感环境数据,按周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)的时空变化及与主要环境因子之间的关系。结果表明,南太平洋长鳍金枪鱼渔场重心主要集中在13°S^23°S,164°E^174°E,总体走势为先向东南方向迁移,后又沿西北方向推移。长鳍金枪鱼全年均可作业,其中5月至7月渔获量最高。作业渔场最适温度(Sea Surface Temperature,SST)范围为26~29℃,相应区域内最适叶绿素a浓度(Chlorophyll a concentration,Chl-a)为0.02~0.05 mg/m3,适宜海面高度距平(Sea surface height anomaly,SSHA)范围为4~24 cm。展开更多
文摘文章利用2008—2015年南太平洋长鳍金枪鱼(Thunnus alalunga)延绳钓渔业数据,结合11个环境指标(海表温度、叶绿素a(Chl-a)浓度、海表温度距平、叶绿素距平、海表温度梯度、叶绿素梯度、海平面异常以及渔区格网对应的前后各1个月海表温度和叶绿素值)和3个时空指标(月、经度和纬度),并基于6种集成学习模型,以月为时间分辨率、0.5°×0.5°为空间分辨率,开展了南太平洋长鳍金枪鱼渔场模型构建和预报研究。模型通过10折交叉验证和网格搜索思想确定最佳参数,采用的随机森林、Bagging决策树、C5.0决策树、梯度提升决策树、AdaBoost、Stacking集成模型分别取得了75.52%、73.87%、72.99%、71.14%、71.33%、75.84%的分类准确率。经对比,Stacking集成模型准确率最高。利用2015年环境数据进行预报精度检验,预报总体准确率为63.86%~82.14%,平均70.99%;高单位捕捞努力量渔获量(Catch per unit effort,CPUE)渔区预报准确率为62.71%~97.85%,平均78.76%。结果表明Stacking集成模型对南太平洋长鳍金枪鱼渔场的预报具有较好的效果及可行性。
基金supported by the National Basic Research Program of China(Grant Nos.2010CB950400 and 2013CB956203)
文摘The characteristic changes of South Pacific sea surface temperature anomalies （SSTAs） for the period January 1979 to December 2011, during which the 1990s Pacific pan-decadal variability （PDV） interdecadal regime shifts occurred, were examined. Empirical Orthogonal Function （EOF） analysis was applied to the monthly mean SSTA for two sub-periods： January 1979 to December 1994 （P 1） and January 1996 to December 2011 （P2）. Both the spatial and temporal features of the leading EOF mode for P1 and P2 showed a remarkable difference. The spatial structure of the leading EOF changed from a tripolar pattern for P 1 （EOF-P 1） to a dipole-like pattern for P2 （EOF-P2）. Besides, EOF-P 1 （EOF-P2） had significant spectral peaks at 4.6 yr （2.7 yr）. EOF-P2 not only had a closer association with E1 Nifio-Southern Oscillation （ENSO）, but also showed a faster response to ENSO than EOF-P1 based on their lead-lag relationships with ENSO. During the development of ENSO, the South Pacific SSTA associated with ENSO for both PI and P2 showed a significant eastward propagation. However, after the peak of ENSO, EOF-P1 showed a stronger persistence than EOF-P2, which still showed eastward propagation. The variability of the SSTA associated with the whole process of ENSO evolution during P1 and the SSTA associated with the development of ENSO during P2 support the existence of ocean-to-atmosphere forcing, but the SSTA associated with the decay of ENSO shows the phenomenon of atmosphere-to-ocean forcing.
文摘长鳍金枪鱼(Thunnus alalung)是南太平洋金枪鱼渔业的主要捕捞种类之一,具有很高的经济价值,掌握其渔场分布与主要海洋环境因子关系是高效开发利用的基础。根据2011年我国南太平洋渔业生产数据和遥感环境数据,按周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)的时空变化及与主要环境因子之间的关系。结果表明,南太平洋长鳍金枪鱼渔场重心主要集中在13°S^23°S,164°E^174°E,总体走势为先向东南方向迁移,后又沿西北方向推移。长鳍金枪鱼全年均可作业,其中5月至7月渔获量最高。作业渔场最适温度(Sea Surface Temperature,SST)范围为26~29℃,相应区域内最适叶绿素a浓度(Chlorophyll a concentration,Chl-a)为0.02~0.05 mg/m3,适宜海面高度距平(Sea surface height anomaly,SSHA)范围为4~24 cm。
