基于残差网络的海洋温跃层分析方法

首先, 以世界海洋地图集2013（WOA13）海洋数据为实验数据, 提出将不等距微分法、 垂直梯度法应用于海洋数据预处理、 海洋区域划分和跃层分析中, 并通过对多种神经网络在基于WOA13海洋三维数据二分类实验的性能分析, 选取残差网络作为二分类实验的网络模型, 在三层残差网络模型基础上增加了Dropout保留层以防止过拟合. 其次, 将残差网络模型用于温跃层分析判定, 并针对改进模型进行超参数优化、 残差单元改进、 保留率调整等对比实验. 实验结果表明, 改进的ResNet 26网络对WOA13海洋区域数据的温跃层数据分类有效, 分类准确率超过94%.     

Evidence for thermocline ventilation in the South China Sea in winter

Based on analysis of Levitus data in the South China Sea (SCS), this note addresses the seasonal feature of the mixed layer and thermocline in the northern SCS in winter, and demonstrates thermocline ventilation during the winter period. The ventilated thermocline is isopycnic layers between sigma 22.0 and 23.5, the other layers beneath is not ventilated yet. The process of thermocline ventilation is accompanied by detrainment water from the mixed layer into the thermocline at the outcropping line and moves to the south. Forced by subduction, the seasonal southward propagating signal also appears in the unventilated thermocline. The horizontal component of the signal propagatesalong the isoline of potential vorticity (PV), in other words, moves around a seasonal PV pool.     

千岛湖极端水位变化对温跃层的影响

为分析千岛湖极端水位变化对温跃层的影响,在水温垂直剖面数据缺测情况下,采用率定后的一维水动力模型DYRESM模拟了不同水位变幅下的温跃层,并设置了极端水位变化情况(情景1)和正常水文条件(情景2)两种情景进行对比分析;基于湖泊分析程序Lake Analyzer计算了两种情景下温跃层表层和底层深度、温跃层厚度和施密特稳定度等。模拟结果表明:情景1水体混合期为2013年11月20日至2014年3月23日,同时段内情景2水体混合期则为2013年12月5日至2014年3月16日;情景1在高水位时,湖体从上到下呈“混合层、温跃层、均温层(30～40m水层)、温跃层、均温层”的双温跃层结构,同时段内情景2则呈“混合层、温跃层、均温层”的单温跃层结构;情景1与情景2相比,平均水位低2.24m,施密特稳定度均值低119.19J/m2,可见极端水位变化条件下出现的低水位情况能显著降低水体稳定度。     

千岛湖水温垂直分布的季节演变对透明溞昼夜垂直迁移的影响

 千岛湖位于钱塘江上游新安江主流上,是中国北亚热带地区的大型深水水体,水温具有明显的分层现象,夏季出现温跃层。枝角类(Cladocera)是水库生态系统的重要类群,具有垂直分布及昼夜迁移的特性。为阐明枝角类垂直分布格局及昼夜迁移对水温垂直分布的生态响应,选择千岛湖坝前段超过60m水深的湖区于2004年9月至2005年8月按季节进行昼夜垂直分层采集,以透明溞(Daphnia hyaline)为对象开展相关研究。结果显示,透明溞是千岛湖坝前段湖区四季的优势种,基本以聚合的模式(MI>1)分布于垂直水柱中,春、夏、冬3季有正常迁移现象,秋季则无明显的昼夜垂直移动行为。千岛湖温跃层的出现加强了透明溞在垂直水柱中的聚集,促进了其夜间的上迁运动,使其成为坝前段湖区表底层水体物质交流的纽带。     

舟山以东浅水海域夏季水团及温盐日变化特征

为了探索舟山以东浅水海域夏季水团分布及温、盐日变化规律,基于舟山以东浅水区域时间分辨率为1h的温盐深(CTD)站点数据,首先对数据剖面取平均以研究其温、盐场结构特征,然后利用Fuzzy聚类法对该区域进行水团划分,最后根据水团划分结果分区域统计分析跃层特征,并计算温、盐日变幅极差值以研究其日变化空间的分布特征。结果表明:研究区域可分为沿岸流水团、近岸混合水团、黑潮水团、东海中-深层水团、东海中-深层混合水团;受黑潮支流与沿岸流混合下沉的影响,调查海域中部中-深层水体相对高温、低盐;沿岸流区与黑潮水区10~20m深度层的温、盐日变幅较大,与跃层深度对应较好,表层、次表层的较小,底层的近似为0;近岸混合水团跃层上界深度最大、平均强度最小,该区域表层、底层日变幅相对较大。     

高温熔融盐单罐蓄热模型与热性能研究

分析了高温熔融盐作为传热流体的单罐斜温层填充床式蓄热器的性能.建立了斜温层单罐蓄热系统的理论模型,并通过数值模拟计算揭示了其蓄热规律.根据系统蓄热量、蓄热能力、熵产和蓄热有效性等确定了评价蓄热器性能的方法,并分析了不同填充材料的物性对斜温层单罐蓄热系统性能的影响.结果显示,在一定的蓄热条件下,较大的导热系数不利于提高蓄...     

南海海洋温差能发电站冷却水排放对周边海洋环境的影响分析

南海温差能资源储量巨大且水深、海底地形条件适宜,是具备开发海洋温差能条件的理想场所。为了分析西沙甘泉岛拟建岛基式海洋温差能发电站冷却水排放对甘泉岛附近海域环境的影响,利用数值模拟的方法建立了三维水动力模型,采用永兴岛验潮站的实测潮位数据验证了模型的可靠性,模拟分析了海洋温差能发电站冷却水的扩散过程及其对甘泉岛附近海域海水水温结构的影响。研究结果表明,温差能发电站冷却水会引起排放口周边温度的急剧下降,进而会在局部范围内产生温跃层,但温降范围主要集中在排水口附近,并且随着潮流流向扩散;排水口处温度梯度最大,且距离排水口越远,温度梯度越小;温跃层包络面积随排水温度的升高而减小,随排水流量的增大而增大;冷水团上方出现一个正跃层,下方则出现一逆跃层,且温跃层上界深度随着排水深度的增大而增大。研究成果可为海洋温差能发电站建设所涉及的环境问题给予合理评价和在生态环境管理上提供科学依据。     

基于自适应阈值的AUV完全自主温跃层探测方法研究

海洋温跃层是一种重要的海洋现象。本文的研究基于自主水下机器人（Autonomous Underwater Vehicle,AUV）,提出了一种梯度阈值可调的温跃层探测方法,其阈值可根据海洋温跃层分布特性自适应修改,同时针对AUV温跃层探测中水体分层厚度选取的问题,提出了一种根据温度梯度峰值点个数自主确定其大小方法,可实现AUV完全自主探测温跃层。根据AUV外场实验数据,把本文方法同峰值法、平均梯度法以及最优分割法进行了比较分析,结果表明该方法简单有效。     

The barrier layer in the southern region of the South China Sea

By analysing the CTD data in the southernregion of the South China Sea gathered during six cruisesbetween 1989 and 1999, a barrier layer with seasonalvariation just like what exists in the equatorial oceans isfound in this region. It is the first discovery in such amarginal sea yet. It is strong in autumn and a little weak in summer and winter. The thicker the barrier layer, the higher the average temperature of the upper mixed layer. The region with the thicker barrier layer overlaps the region with thehigher average temperature of the upper mixed layer, andaccords with the thicker region of the warm pool in the South China Sea got from the Levitus data. The barrier layer in the southern region of the South China Sea has significantinfluence on the heat storage of the upper ocean there.     

Influences of the surface wave-induced mixing and tidal mixing on the vertical temperature structure of the Yellow and East China Seas in summer

After a classification of the physical processes which affect the vertical mixing, diffusivity induced by the surface wave momentum and tidal currents and its influence on the vertical temperature structure are discussed. Based on three mixing schemes, the vertical temperature structure of the Yellow and East China Seas (YECS) is simulated. Results show that in summer, the surface wave-induced mixing plays a key role in forming the upper mixed layer in the YECS. The tidal mixing controls the lower layers within 30 m above the bottom, which is the main factor in forming the platform-shaped temperature structure in the southern Yellow Sea (YS). Together with the strong surface wave-induced mixing, the tidal mixing makes the thermocline ventilate near the east coast of the southern YS. The double cold cores in the deeper layers of the East China Sea have different causes. The western one is the maintenance of the winter cold water, while the eastern one is configured by circulation. The simulated vertical temperature structure of the YECS with the surface wave-induced mixing and tidal mixing has similar features to the observations, which indicates that these mixing processes are key factors in simulating the thermocline and pycnocline of the coastal oceans.