期刊文献+
共找到2篇文章
< 1 >
每页显示 20 50 100
深度学习在印度洋偶极子预报中的应用研究 被引量:3
1
作者 刘俊 唐佑民 +1 位作者 宋迅殊 孙志林 《大气科学》 CSCD 北大核心 2022年第3期590-598,共9页
印度洋偶极子(IOD)是热带印度洋秋季最强的年际变率,它会通过大气遥相关来影响世界许多地区的气候。目前耦合气候模式对IOD预报技巧仍非常有限,远低于热带太平洋的厄尔尼诺事件的预报技巧。鉴于深度学习具备高效的数据处理能力,本文使... 印度洋偶极子(IOD)是热带印度洋秋季最强的年际变率,它会通过大气遥相关来影响世界许多地区的气候。目前耦合气候模式对IOD预报技巧仍非常有限,远低于热带太平洋的厄尔尼诺事件的预报技巧。鉴于深度学习具备高效的数据处理能力,本文使用深度学习中的卷积神经网络(CNN)与人工神经网络中的多层感知机(MLP)处理再分析观测资料,从而进行IOD预报。由于当预报初始时刻为北半球冬春季时,对IOD事件的预报技巧较低。因此,为探索CNN的预报能力,本文仅使用三种(1~3月、2~4月、3~5月)初始时刻的海表温度异常(SSTA)作为CNN的输入数据,来预报其后续七个月的印度洋偶极子指数(DMI)、东极子指数(EIOI)和西极子指数(WIOI)。结果表明:CNN对DMI、EIOI和WIOI的有效预测时效均超过了6个月。与现在耦合动力模式相比,CNN模型能够显著提升DMI和EIOI的预报技巧,但对WIOI的预报技巧提升有限。当预报提前时间为7个月时,CNN模型能够比较准确地预报1994、1997与2019年的IOD事件。由于CNN模型能够更好地抓住印度洋海温的空间结构特征,它对IOD事件的预报技巧比传统神经网络MLP高。 展开更多
关键词 印度洋偶极子 深度学习 卷积神经网络 气候预报
下载PDF
The distribution of chlorophyll a in the tropical eastern Indian Ocean in austral summer 被引量:3
2
作者 HONG Lisha WANG Chunsheng +4 位作者 ZHOU Yadong CHEN Mianrun LIU Hongbin LIN Zhongyang song xunshu 《Acta Oceanologica Sinica》 SCIE CAS CSCD 2012年第5期146-159,共14页
To study the effect of hydrographic factors on the spatial distributions of chlorophyll a (Chl a), an investigation was carried out in the tropical eastern Indian Ocean (80 –100 E along 7 S, and 7 –18 S along 80 ... To study the effect of hydrographic factors on the spatial distributions of chlorophyll a (Chl a), an investigation was carried out in the tropical eastern Indian Ocean (80 –100 E along 7 S, and 7 –18 S along 80 E) in December 2010. The fluorescent method was used to obtain total Chl a and size-fractioned Chl a at the 26 stations. The results show that surface Chl a concentration averaged at (0.168 ± 0.095) mg/m 3 s.d. (range: 0.034–0.475 mg/m 3 ), concentrations appeared to be higher in the west for longitudinal variations, and higher in the north for latitudinal variations. Furthermore, the surface Chl a concentration was lower (0.034–0.066 mg/m 3 ) in the region to the south of 16 S. There was a strong subsurface Chl a maximum layer at all stations and the depth of the Chl a maximum increased towards to the east and south along with the respective nitracline. The spatial variation of Chl a was significant: correlation and regression analysis suggests that it was primarily affected by PO 3 4 , N(NO 3 –N+NO 2 –N) and temperature. Size-fractionated Chl a concentration clearly showed that the study area was a typical oligotrophic open ocean, in which picophytoplankton dominated, accounting for approximately 67.8% of total Chl a, followed by nanophytoplankton (24.5%) and microphytoplankton (7.6%). The two larger fractions were sensitive to the limitation of P, while picophytoplankton was primarily affected by temperature. 展开更多
关键词 eastern Indian Ocean chlorophyll a size fraction Indonesian Throughflow
下载PDF
上一页 1 下一页 到第
使用帮助 返回顶部