Ocean Modelling最新文献_第10页

Wave dynamics in the Yellow River Estuary during cold wave and typhoon events 寒潮和台风期间黄河口的波浪动力学

IF 3.1 3区地球科学 Q2 METEOROLOGY & ATMOSPHERIC SCIENCES

Ocean Modelling

Pub Date : 2025-09-01 Epub Date: 2025-05-30 DOI: 10.1016/j.ocemod.2025.102568

Shenghan Gao , Miaohua Mao , Meng Xia

The Yellow River Estuary (YRE), located in the semi-enclosed Bohai Sea, is characterized by complex shorelines and shallow water depths and is vulnerable to high waves during extreme weather events. Therefore, a dual-nested third-generation wave model was applied to investigate wave dynamics during Typhoons In-Fa (2021) and Mui-Fa (2022) and a pair of cold wave events in 2021 and 2022. The YRE model was refined to reproduce realistic high-resolution terrain and then calibrated against observations at four long-term buoy stations. Results indicate that wave characteristics closely correlate with winds, modulated by local bathymetry. During cold waves, the temporal evolution of the significant wave height (H_s) exhibits double peaks, whereas a single peak is observed during typhoons due to alternative development responses to winds. This resulted in a 1.5-h time lag between H_s and winds. Wind waves primarily dominate sea states, while swells occur after the typhoon passage. Bathymetric refraction plays an essential role in sheltering the southern region of the YRE from remotely energetic swells. Further investigations reveal that depth-induced breaking and whitecapping jointly control wave energy dissipation. Bathymetric heterogeneity and shoaling processes substantially influence wave energy, resulting in wave attenuation and spatial variability. Intense triad wave-wave interactions and wave breaking contribute to increased H_s, causing multiple wave-breaking processes during propagation. The findings in the YRE help enhance the understanding of wave dynamics in similar shallow-water mega deltas and estuaries.

黄河口位于半封闭的渤海海域，岸线复杂，水深浅，在极端天气条件下易受大浪影响。因此，采用双嵌套第三代波浪模型对台风“台风”（2021）和台风“台风”（2022）以及2021年和2022年两次寒潮事件的波浪动力学进行了研究。YRE模型经过改进以再现真实的高分辨率地形，然后根据四个长期浮标站的观测结果进行校准。结果表明，波浪特征与风密切相关，受局部测深调制。在寒潮期间，有效波高（Hs）的时间演变呈现双峰，而在台风期间，由于对风的替代发展响应，观测到单峰。这导致h和风之间有1.5小时的时间差。风浪主要主导海况，而巨浪则在台风通过后出现。深海折射在保护YRE南部地区免受遥远的能量膨胀方面起着至关重要的作用。进一步的研究表明，深度破碎和白浪共同控制着波浪的能量耗散。水深非均质性和浅滩化过程对波浪能量有很大影响，导致波浪衰减和空间变异性。强烈的三联波相互作用和破波有助于增加Hs，在传播过程中引起多次破波过程。YRE的发现有助于提高对类似的浅水巨型三角洲和河口波浪动力学的理解。

{"title":"Wave dynamics in the Yellow River Estuary during cold wave and typhoon events","authors":"Shenghan Gao , Miaohua Mao , Meng Xia","doi":"10.1016/j.ocemod.2025.102568","DOIUrl":"10.1016/j.ocemod.2025.102568","url":null,"abstract":"<div><div>The Yellow River Estuary (YRE), located in the semi-enclosed Bohai Sea, is characterized by complex shorelines and shallow water depths and is vulnerable to high waves during extreme weather events. Therefore, a dual-nested third-generation wave model was applied to investigate wave dynamics during Typhoons In-Fa (2021) and Mui-Fa (2022) and a pair of cold wave events in 2021 and 2022. The YRE model was refined to reproduce realistic high-resolution terrain and then calibrated against observations at four long-term buoy stations. Results indicate that wave characteristics closely correlate with winds, modulated by local bathymetry. During cold waves, the temporal evolution of the significant wave height (<em>H<sub>s</sub></em>) exhibits double peaks, whereas a single peak is observed during typhoons due to alternative development responses to winds. This resulted in a 1.5-h time lag between <em>H<sub>s</sub></em> and winds. Wind waves primarily dominate sea states, while swells occur after the typhoon passage. Bathymetric refraction plays an essential role in sheltering the southern region of the YRE from remotely energetic swells. Further investigations reveal that depth-induced breaking and whitecapping jointly control wave energy dissipation. Bathymetric heterogeneity and shoaling processes substantially influence wave energy, resulting in wave attenuation and spatial variability. Intense triad wave-wave interactions and wave breaking contribute to increased <em>H<sub>s</sub></em>, causing multiple wave-breaking processes during propagation. The findings in the YRE help enhance the understanding of wave dynamics in similar shallow-water mega deltas and estuaries.</div></div>","PeriodicalId":19457,"journal":{"name":"Ocean Modelling","volume":"197 ","pages":"Article 102568"},"PeriodicalIF":3.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144212922","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Revisiting tidal rectification by bottom topography 从底部地形再看潮汐整流

IF 3.1 3区地球科学 Q2 METEOROLOGY & ATMOSPHERIC SCIENCES

Ocean Modelling

Pub Date : 2025-09-01 Epub Date: 2025-07-15 DOI: 10.1016/j.ocemod.2025.102587

Logueminda Sabaga , Yves Morel , Nadia Ayoub , Patrick Marsaleix , Hoavo Hova , Alexis Chaigneau

Tidal rectification plays a key role in controlling mean transport in coastal areas and coast-basin material exchange. To calculate mean flows, conventional approaches require high-resolution basin-scale numerical simulations which demands substantial computational resources. This study revisits tidal rectification governed by topographic variation and bottom friction, and proposes a new analytical solution.

The first step is to derive solutions in the simplest possible configuration. We thus revisit solutions in one-dimensional (1D) configurations, using a Lagrangian approach from which Eulerian results are derived. Exact solutions are provided for the frictionless case and new approximate solutions are developed for a more realistic quadratic bottom friction.

We then analyze the influence of viscosity on solutions from numerical models. We find that the latter has moderate influence when quadratic bottom friction is considered. However, when the steady rectified current extends over regions deeper than a critical depth, viscosity can lead to spurious effects and alter the accuracy of the numerical results. We show the critical depth can be expressed as a function of friction coefficient, tidal flux and topography variation length-scale.

We finally extend the analytical solutions derived for the 1D case to the two-dimensional (2D) case. The 2D solutions are compared to results from an ocean general circulation model solving the full barotropic equations in an academic configuration with a complex topography and a quadratic bottom friction. Comparison between analytical solutions and numerical simulations shows good agreement for both the magnitude and direction of the steady rectified tidal current. Sensitivity tests to bottom friction and tide amplitude show that the steady rectified current is parallel to the isobaths and independent of the magnitude of the bottom friction coefficient at first order.

潮汐整流在控制沿海地区平均输运和海岸-盆地物质交换中起着关键作用。为了计算平均流量，传统的方法需要高分辨率的流域尺度数值模拟，这需要大量的计算资源。本文重新研究了地形变化和海底摩擦对潮汐整流的影响，并提出了一种新的解析解。第一步是在尽可能简单的配置中推导解。因此，我们重新审视一维（1D）构型的解决方案，使用拉格朗日方法，从欧拉结果推导。给出了无摩擦情况下的精确解，并给出了更为实际的二次底摩擦情况下的近似解。然后通过数值模型分析了粘度对溶液的影响。当考虑二次底摩擦时，后者的影响较小。然而，当稳定整流电流延伸到超过临界深度的区域时，粘度会导致虚假效应并改变数值结果的准确性。结果表明，临界深度可以表示为摩擦系数、潮汐通量和地形变化长度尺度的函数。最后，我们将一维情况下的解析解扩展到二维（2D）情况。将二维解与海洋环流模型在复杂地形和二次底摩擦条件下求解全正压方程的结果进行了比较。解析解与数值模拟结果的比较表明，稳态整流潮流的大小和方向符合较好。对底摩擦和潮汐幅值的敏感性试验表明，稳定整流电流平行于等深线，与一阶底摩擦系数的大小无关。

{"title":"Revisiting tidal rectification by bottom topography","authors":"Logueminda Sabaga , Yves Morel , Nadia Ayoub , Patrick Marsaleix , Hoavo Hova , Alexis Chaigneau","doi":"10.1016/j.ocemod.2025.102587","DOIUrl":"10.1016/j.ocemod.2025.102587","url":null,"abstract":"<div><div>Tidal rectification plays a key role in controlling mean transport in coastal areas and coast-basin material exchange. To calculate mean flows, conventional approaches require high-resolution basin-scale numerical simulations which demands substantial computational resources. This study revisits tidal rectification governed by topographic variation and bottom friction, and proposes a new analytical solution.</div><div>The first step is to derive solutions in the simplest possible configuration. We thus revisit solutions in one-dimensional (1D) configurations, using a Lagrangian approach from which Eulerian results are derived. Exact solutions are provided for the frictionless case and new approximate solutions are developed for a more realistic quadratic bottom friction.</div><div>We then analyze the influence of viscosity on solutions from numerical models. We find that the latter has moderate influence when quadratic bottom friction is considered. However, when the steady rectified current extends over regions deeper than a critical depth, viscosity can lead to spurious effects and alter the accuracy of the numerical results. We show the critical depth can be expressed as a function of friction coefficient, tidal flux and topography variation length-scale.</div><div>We finally extend the analytical solutions derived for the 1D case to the two-dimensional (2D) case. The 2D solutions are compared to results from an ocean general circulation model solving the full barotropic equations in an academic configuration with a complex topography and a quadratic bottom friction. Comparison between analytical solutions and numerical simulations shows good agreement for both the magnitude and direction of the steady rectified tidal current. Sensitivity tests to bottom friction and tide amplitude show that the steady rectified current is parallel to the isobaths and independent of the magnitude of the bottom friction coefficient at first order.</div></div>","PeriodicalId":19457,"journal":{"name":"Ocean Modelling","volume":"197 ","pages":"Article 102587"},"PeriodicalIF":3.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144672651","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Developing Intelligent Earth System Models : An AI scheme of K-profile parameterization and stable coupling into CESM with FTA 开发智能地球系统模型：一种k -廓线参数化和与FTA稳定耦合的人工智能方案

IF 3.1 3区地球科学 Q2 METEOROLOGY & ATMOSPHERIC SCIENCES

Ocean Modelling

Pub Date : 2025-09-01 Epub Date: 2025-06-11 DOI: 10.1016/j.ocemod.2025.102567

Bin Mu , Kang Yang , Bo Qin , Hao Li , Shijin Yuan

Parameterization schemes in numerical models are employed to represent the effects of subgrid-scale physical processes but are often limited by incomplete understanding of physical processes and computational constraints, leading to inaccuracies and inefficiencies. Artificial intelligence (AI) models have been introduced to enhance simulation accuracy or computational efficiency. However, hybrid Earth System Models (ESMs), which integrate AI into traditional frameworks, must also consider stability in coupled simulations. In this study, we replace the default K-profile parameterization (KPP) in Community Earth System Model (CESM) with a transformer-based AI model (KPP-DL). We first perform offline evaluations, demonstrating the AI model’s ability to closely replicate KPP’s key outputs. Subsequently, we couple KPP-DL into CESM via Fortran-Torch adaptor (FTA) and evaluate the hybrid CESM’s performance in terms of accuracy, stability, and computational efficiency. Hybrid CESM maintains stable operation for at least 3 years, with approximately a 3-5 times improvement in the computational efficiency of vertical mixing. During online coupled simulations, KPP-DL exhibits strong agreement with KPP in simulating key vertical mixing coefficients while hybrid CESM produces consistent results for variables such as temperature and salinity. Our results highlight the potential of AI-driven approaches to achieve accuracy approaching that of KPP and stability coupled in ESMs while improving the efficiency, suggesting that intelligent ESMs represent a promising future direction for numerical modeling.

数值模型中的参数化方案被用来表示亚网格尺度物理过程的影响，但往往受到对物理过程的不完全理解和计算约束的限制，导致不准确和低效率。人工智能（AI）模型被引入以提高仿真精度或计算效率。然而，将人工智能集成到传统框架中的混合地球系统模型（ESMs）也必须考虑耦合模拟的稳定性。在本研究中，我们用基于变压器的人工智能模型（KPP- dl）取代了社区地球系统模型（CESM）中默认的k -剖面参数化（KPP）。我们首先进行离线评估，展示了人工智能模型密切复制KPP关键输出的能力。随后，我们通过Fortran-Torch适配器（FTA）将KPP-DL耦合到CESM中，并从精度、稳定性和计算效率方面评估混合CESM的性能。混合CESM保持稳定运行至少3年，垂直混合的计算效率提高了约3-5倍。在在线耦合模拟中，KPP- dl与KPP在模拟关键垂直混合系数方面表现出很强的一致性，而混合CESM对温度和盐度等变量的模拟结果一致。我们的研究结果强调了人工智能驱动的方法在提高效率的同时实现接近KPP和稳定性耦合在esm中的精度的潜力，这表明智能esm代表了数值模拟的一个有希望的未来方向。

{"title":"Developing Intelligent Earth System Models : An AI scheme of K-profile parameterization and stable coupling into CESM with FTA","authors":"Bin Mu , Kang Yang , Bo Qin , Hao Li , Shijin Yuan","doi":"10.1016/j.ocemod.2025.102567","DOIUrl":"10.1016/j.ocemod.2025.102567","url":null,"abstract":"<div><div>Parameterization schemes in numerical models are employed to represent the effects of subgrid-scale physical processes but are often limited by incomplete understanding of physical processes and computational constraints, leading to inaccuracies and inefficiencies. Artificial intelligence (AI) models have been introduced to enhance simulation accuracy or computational efficiency. However, hybrid Earth System Models (ESMs), which integrate AI into traditional frameworks, must also consider stability in coupled simulations. In this study, we replace the default K-profile parameterization (KPP) in Community Earth System Model (CESM) with a transformer-based AI model (KPP-DL). We first perform offline evaluations, demonstrating the AI model’s ability to closely replicate KPP’s key outputs. Subsequently, we couple KPP-DL into CESM via Fortran-Torch adaptor (FTA) and evaluate the hybrid CESM’s performance in terms of accuracy, stability, and computational efficiency. Hybrid CESM maintains stable operation for at least 3 years, with approximately a 3-5 times improvement in the computational efficiency of vertical mixing. During online coupled simulations, KPP-DL exhibits strong agreement with KPP in simulating key vertical mixing coefficients while hybrid CESM produces consistent results for variables such as temperature and salinity. Our results highlight the potential of AI-driven approaches to achieve accuracy approaching that of KPP and stability coupled in ESMs while improving the efficiency, suggesting that intelligent ESMs represent a promising future direction for numerical modeling.</div></div>","PeriodicalId":19457,"journal":{"name":"Ocean Modelling","volume":"197 ","pages":"Article 102567"},"PeriodicalIF":3.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144298194","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Waves in a temperate, microtidal and restricted Mediterranean coastal lagoon 在温带、微潮和受限制的地中海沿岸泻湖中的波浪

IF 3.1 3区地球科学 Q2 METEOROLOGY & ATMOSPHERIC SCIENCES

Ocean Modelling

Pub Date : 2025-09-01 Epub Date: 2025-06-10 DOI: 10.1016/j.ocemod.2025.102578

Bartolomé Morote-Sánchez , Francisco López-Castejón , Javier Gilabert

The study focused on wave dynamics in the Mar Menor, a hypersaline coastal lagoon in the Southwestern Mediterranean Sea, prompted by ecological crises and the need to understand the physical drivers of its ecology, especially sediment transport and nutrient resuspension. The research employs the SWAN model for wave simulations forced with recorded winds from a meteorological station in the lagoon, covering data from August to October 2019. Model results were validated against data from an ADCP deployed near the met station. The results indicated that Mar Menor experiences low-intensity winds, with occasional strong Northeasterly winds causing the highest waves of up to 1.25 meters high. Self-Organizing Maps (SOMs) analysis provides a classification of wave height, period, wavelength, and bottom orbital velocity, resulting in six wave map categories. The analysis revealed that the largest waves are linked to Southerly winds, and sediment resuspension is most significant during storms, not affecting the deepest 6-meter depth central part of the lagoon. The study concludes that wave-induced orbital velocities can mobilize sediment resuspension during extreme events, potentially disturbing the anoxic bottom. The lagoon was strongly stratified after a flash flood occurred between 12–15 September 2019 with an anoxic bottom layer. Waves driven by winds blowing for long time, although not strongly, contributed to break the stratification producing the fish mass mortality observed in the northern part of the lagoon one month later. Identifying specific wind patterns associated with each SOM map category shows the predictive potential of SOMs for forecasting wave patterns from wind data. Due to their low computational requirements and reliance solely on wind forecasts, SOMs analysis offer a practical tool for early warning systems and for managing ecological risks in environmentally sensitive areas such as coastal lagoons.

由于生态危机和了解其生态的物理驱动因素，特别是沉积物运输和营养再悬浮的需要，这项研究的重点是地中海西南部一个高盐沿海泻湖Mar Menor的波浪动力学。该研究采用SWAN模型对泻湖气象站记录的风进行波浪模拟，涵盖2019年8月至10月的数据。模型结果与部署在气象站附近的ADCP的数据进行了验证。结果表明，马诺岛的风力强度较低，偶有强烈的东北风，浪高最高可达1.25米。自组织图（SOMs）分析提供了波高、周期、波长和底轨道速度的分类，从而产生六种波图类别。分析显示，最大的波浪与南风有关，沉积物再悬浮在风暴期间最为显著，不会影响泻湖中心最深的6米深的部分。该研究的结论是，波浪引起的轨道速度可以在极端事件中调动沉积物的再悬浮，潜在地扰乱缺氧的底部。2019年9月12日至15日发生山洪暴发后，泻湖形成了强烈的分层，底层缺氧。虽然不强烈，但长时间的风驱动的波浪有助于打破分层，造成一个月后在泻湖北部观察到的鱼类大量死亡。识别与每个SOM地图类别相关的特定风型，显示了SOM从风数据预测波浪型的预测潜力。由于SOMs的计算需求低且仅依赖于风力预报，因此它为早期预警系统和管理环境敏感地区（如沿海泻湖）的生态风险提供了实用工具。

{"title":"Waves in a temperate, microtidal and restricted Mediterranean coastal lagoon","authors":"Bartolomé Morote-Sánchez , Francisco López-Castejón , Javier Gilabert","doi":"10.1016/j.ocemod.2025.102578","DOIUrl":"10.1016/j.ocemod.2025.102578","url":null,"abstract":"<div><div>The study focused on wave dynamics in the Mar Menor, a hypersaline coastal lagoon in the Southwestern Mediterranean Sea, prompted by ecological crises and the need to understand the physical drivers of its ecology, especially sediment transport and nutrient resuspension. The research employs the SWAN model for wave simulations forced with recorded winds from a meteorological station in the lagoon, covering data from August to October 2019. Model results were validated against data from an ADCP deployed near the met station. The results indicated that Mar Menor experiences low-intensity winds, with occasional strong Northeasterly winds causing the highest waves of up to 1.25 meters high. Self-Organizing Maps (SOMs) analysis provides a classification of wave height, period, wavelength, and bottom orbital velocity, resulting in six wave map categories. The analysis revealed that the largest waves are linked to Southerly winds, and sediment resuspension is most significant during storms, not affecting the deepest 6-meter depth central part of the lagoon. The study concludes that wave-induced orbital velocities can mobilize sediment resuspension during extreme events, potentially disturbing the anoxic bottom. The lagoon was strongly stratified after a flash flood occurred between 12–15 September 2019 with an anoxic bottom layer. Waves driven by winds blowing for long time, although not strongly, contributed to break the stratification producing the fish mass mortality observed in the northern part of the lagoon one month later. Identifying specific wind patterns associated with each SOM map category shows the predictive potential of SOMs for forecasting wave patterns from wind data. Due to their low computational requirements and reliance solely on wind forecasts, SOMs analysis offer a practical tool for early warning systems and for managing ecological risks in environmentally sensitive areas such as coastal lagoons.</div></div>","PeriodicalId":19457,"journal":{"name":"Ocean Modelling","volume":"197 ","pages":"Article 102578"},"PeriodicalIF":3.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144298196","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Wavelet ocean data assimilation 小波海洋资料同化

IF 3.1 3区地球科学 Q2 METEOROLOGY & ATMOSPHERIC SCIENCES

Ocean Modelling

Pub Date : 2025-09-01 Epub Date: 2025-07-02 DOI: 10.1016/j.ocemod.2025.102589

Bradley Sciacca , Hans Ngodock , Joseph M. D’Addezio , Matthew J. Carrier , Innocent Souopgui

Due to necessary assumptions of observational errors with an exigency for appropriate and timely inversion in the assimilation, dense observations are thinned and/or altered before being assimilated into ocean models. Historically, this process did not significantly restrict model skill because most of the observation types had a quite coarse horizontal distribution. However, recent advances in observation resolution demand new assimilation approaches, whereby small-scale features are actively corrected in the model background. A novel method is introduced that applies multiscale data assimilation utilizing the wavelet transform. Unlike other currently employed ocean multiscale techniques, this method is performed in a single analysis step. Utilizing the wavelet transform allows for observational information to be retained at all its original grid points, compared to the averaging and removal in traditional techniques, such as super observations. This comes from the unique space and frequency relation available to the wavelet transform, which instead filters the potentially correlated small-scale observation errors at each model grid point. Several six-month identical twin data assimilation experiments were used to validate the method. Results indicate comparable to substantial improvements over super observations. On average, the sea surface temperature RMSE was 39 % lower for the wavelet method over the six-months compared to super observations. The wavelet method was also able to constrain horizontal scales in assimilation 29 km and above compared to 60 km and above for the super observations.

由于对观测误差的必要假设以及同化过程中需要适当和及时的反演，密集观测在被同化到海洋模式之前会被稀释和/或改变。从历史上看，这个过程并没有显著地限制模型技能，因为大多数观测类型具有相当粗糙的水平分布。然而，观测分辨率的最新进展需要新的同化方法，即在模式背景中主动校正小尺度特征。提出了一种利用小波变换进行多尺度数据同化的新方法。与目前使用的其他海洋多尺度技术不同，该方法在单个分析步骤中完成。与传统技术（如超级观测）的平均和去除相比，利用小波变换可以将观测信息保留在所有原始网格点上。这是因为小波变换具有独特的空间和频率关系，而小波变换可以过滤每个模型网格点上潜在的相关小尺度观测误差。用几个六个月的同卵双胞胎数据同化实验来验证该方法。结果表明，与超级观测相比，有相当大的改进。与超级观测值相比，小波法测得的6个月平均海面温度RMSE降低了39%。与60 km及以上的超级观测值相比，小波方法还能够约束同化29 km及以上的水平尺度。

{"title":"Wavelet ocean data assimilation","authors":"Bradley Sciacca , Hans Ngodock , Joseph M. D’Addezio , Matthew J. Carrier , Innocent Souopgui","doi":"10.1016/j.ocemod.2025.102589","DOIUrl":"10.1016/j.ocemod.2025.102589","url":null,"abstract":"<div><div>Due to necessary assumptions of observational errors with an exigency for appropriate and timely inversion in the assimilation, dense observations are thinned and/or altered before being assimilated into ocean models. Historically, this process did not significantly restrict model skill because most of the observation types had a quite coarse horizontal distribution. However, recent advances in observation resolution demand new assimilation approaches, whereby small-scale features are actively corrected in the model background. A novel method is introduced that applies multiscale data assimilation utilizing the wavelet transform. Unlike other currently employed ocean multiscale techniques, this method is performed in a single analysis step. Utilizing the wavelet transform allows for observational information to be retained at all its original grid points, compared to the averaging and removal in traditional techniques, such as super observations. This comes from the unique space and frequency relation available to the wavelet transform, which instead filters the potentially correlated small-scale observation errors at each model grid point. Several six-month identical twin data assimilation experiments were used to validate the method. Results indicate comparable to substantial improvements over super observations. On average, the sea surface temperature RMSE was 39 % lower for the wavelet method over the six-months compared to super observations. The wavelet method was also able to constrain horizontal scales in assimilation 29 km and above compared to 60 km and above for the super observations.</div></div>","PeriodicalId":19457,"journal":{"name":"Ocean Modelling","volume":"197 ","pages":"Article 102589"},"PeriodicalIF":3.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144572677","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A global high-resolution CMIP6 ensemble of wave climate simulations and projections using a coastal multigrid: Configuration and performance evaluation 基于沿海多重网格的全球高分辨率CMIP6海浪气候模拟和预估：配置和性能评估

IF 3.1 3区地球科学 Q2 METEOROLOGY & ATMOSPHERIC SCIENCES

Ocean Modelling

Pub Date : 2025-09-01 Epub Date: 2025-05-29 DOI: 10.1016/j.ocemod.2025.102566

Rajesh Kumar , Gil Lemos , Alvaro Semedo , Jian-Guo Li

This study presents a comprehensive evaluation of a high-resolution wave climate ensemble, driven by eight CMIP6 General Circulation Models (GCMs), using a coastal multigrid approach based on Spherical Multiple-Cell (SMC) grid. The use of the SMC grid allows for global wave climate simulations refined to high resolutions up to 6 km in coastal regions, where complex interactions between wind, waves, and bathymetry demand more precise modelling. The ensemble’s performance is assessed against wave reanalysis datasets and near-coastal in-situ wave observations, with a focus on key wave climate parameters: significant wave height, mean wave period, peak wave period, and mean wave direction. Results show that the ensemble is able to accurately represent the historical wave climate across diverse regions, excelling in coastal areas, when compared with previous similar datasets, namely across Europe, North America and the Maritime Continent. The multi-resolution SMC grid captures complicated coastal wave patterns and improves the representation of coastal wave dynamics. The ensemble’s ability to simulate both seasonal variability and extreme wave events highlights its potential for high-resolution global wave climate projections, with multiple applications for coastal management and adaptation strategies, marking an advancement in wave climate modelling through its integration in high-resolution, multigrid frameworks.

本文利用基于球面多单元格网（SMC）的沿海多网格方法，对8个CMIP6环流模式（GCMs）驱动的高分辨率波浪气候集合进行了综合评价。SMC网格的使用允许在沿海地区将全球波浪气候模拟精细到高达6公里的高分辨率，在这些地区，风、波和水深测量之间的复杂相互作用需要更精确的建模。根据波浪再分析数据集和近岸现场波浪观测对该系统的性能进行了评估，重点关注了关键的波浪气候参数：有效波高、平均波周期、峰值波周期和平均波方向。结果表明，与以往的类似数据集（即欧洲、北美和海洋大陆）相比，该集合能够准确地代表不同地区的历史波浪气候，在沿海地区表现出色。多分辨率SMC网格捕捉了复杂的海岸波型，改善了海岸波动力学的表征。该集合模拟季节变化和极端波浪事件的能力突出了其在高分辨率全球波浪气候预测方面的潜力，在沿海管理和适应策略方面具有多种应用，通过集成高分辨率、多网格框架，标志着波浪气候建模的进步。

{"title":"A global high-resolution CMIP6 ensemble of wave climate simulations and projections using a coastal multigrid: Configuration and performance evaluation","authors":"Rajesh Kumar , Gil Lemos , Alvaro Semedo , Jian-Guo Li","doi":"10.1016/j.ocemod.2025.102566","DOIUrl":"10.1016/j.ocemod.2025.102566","url":null,"abstract":"<div><div>This study presents a comprehensive evaluation of a high-resolution wave climate ensemble, driven by eight CMIP6 General Circulation Models (GCMs), using a coastal multigrid approach based on Spherical Multiple-Cell (SMC) grid. The use of the SMC grid allows for global wave climate simulations refined to high resolutions up to 6 km in coastal regions, where complex interactions between wind, waves, and bathymetry demand more precise modelling. The ensemble’s performance is assessed against wave reanalysis datasets and near-coastal <em>in-situ</em> wave observations, with a focus on key wave climate parameters: significant wave height, mean wave period, peak wave period, and mean wave direction. Results show that the ensemble is able to accurately represent the historical wave climate across diverse regions, excelling in coastal areas, when compared with previous similar datasets, namely across Europe, North America and the Maritime Continent. The multi-resolution SMC grid captures complicated coastal wave patterns and improves the representation of coastal wave dynamics. The ensemble’s ability to simulate both seasonal variability and extreme wave events highlights its potential for high-resolution global wave climate projections, with multiple applications for coastal management and adaptation strategies, marking an advancement in wave climate modelling through its integration in high-resolution, multigrid frameworks.</div></div>","PeriodicalId":19457,"journal":{"name":"Ocean Modelling","volume":"197 ","pages":"Article 102566"},"PeriodicalIF":3.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144231206","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Disentangling wavy and vortical motions in concurrent snapshots of the sea surface height and velocity 在海面高度和速度的同步快照中解开波浪和漩涡运动

IF 3.1 3区地球科学 Q2 METEOROLOGY & ATMOSPHERIC SCIENCES

Ocean Modelling

Pub Date : 2025-07-01 Epub Date: 2025-04-18 DOI: 10.1016/j.ocemod.2025.102556

Chuanyin Wang , Zhiyu Liu , Hongyang Lin , Cesar Rocha , Qinghua Yang , Dake Chen , Junbin Gong

Wide-swath satellite missions, such as Surface Water and Ocean Topography (SWOT) and Ocean Dynamics and Sea Exchanges with the Atmosphere (ODYSEA), will provide quasi-concurrent observations of the two-dimensional sea surface height and velocity. Thanks to their high spatial resolution, the spatial features of both vortical and wavy oceanic motions are expected to be captured by these observations. A natural question is whether one can disentangle vortical and wavy motions in these snapshot observations. This issue has attracted some efforts, but crucial progress remains to be made. Here, assuming that only a single concurrent snapshot of the sea surface height and velocity is available, we pursue a dynamical approach for disentangling vortical and wavy motions. This is realized by noting that wavy motions do not induce potential vorticity anomalies. A proof-of-concept application using an output of a realistic high-resolution numerical simulation suggests that the proposed approach is simple and efficient, and is particularly useful for separating wavy and vortical motions in observations by wide-swath satellite missions.

大范围卫星任务，如地表水和海洋地形（SWOT）和海洋动力学和海洋与大气交换（ODYSEA），将提供二维海面高度和速度的准同步观测。由于它们的高空间分辨率，预计这些观测将捕捉到涡旋和波浪状海洋运动的空间特征。一个自然的问题是，人们是否可以在这些快照观测中区分出漩涡运动和波浪运动。这个问题已引起一些努力，但仍有待取得关键进展。在这里，假设只有海面高度和速度的单一并发快照是可用的，我们追求一个动态的方法来解开漩涡和波浪运动。这是通过注意到波浪运动不会引起潜在涡度异常而实现的。一项使用实际高分辨率数值模拟输出的概念验证应用表明，所提出的方法简单有效，特别适用于在大范围卫星任务观测中分离波浪运动和涡旋运动。

{"title":"Disentangling wavy and vortical motions in concurrent snapshots of the sea surface height and velocity","authors":"Chuanyin Wang , Zhiyu Liu , Hongyang Lin , Cesar Rocha , Qinghua Yang , Dake Chen , Junbin Gong","doi":"10.1016/j.ocemod.2025.102556","DOIUrl":"10.1016/j.ocemod.2025.102556","url":null,"abstract":"<div><div>Wide-swath satellite missions, such as Surface Water and Ocean Topography (SWOT) and Ocean Dynamics and Sea Exchanges with the Atmosphere (ODYSEA), will provide quasi-concurrent observations of the two-dimensional sea surface height and velocity. Thanks to their high spatial resolution, the spatial features of both vortical and wavy oceanic motions are expected to be captured by these observations. A natural question is whether one can disentangle vortical and wavy motions in these snapshot observations. This issue has attracted some efforts, but crucial progress remains to be made. Here, assuming that only a single concurrent snapshot of the sea surface height and velocity is available, we pursue a dynamical approach for disentangling vortical and wavy motions. This is realized by noting that wavy motions do not induce potential vorticity anomalies. A proof-of-concept application using an output of a realistic high-resolution numerical simulation suggests that the proposed approach is simple and efficient, and is particularly useful for separating wavy and vortical motions in observations by wide-swath satellite missions.</div></div>","PeriodicalId":19457,"journal":{"name":"Ocean Modelling","volume":"196 ","pages":"Article 102556"},"PeriodicalIF":3.1,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143855274","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

APPLE-MASNUM: Accelerating parallel processing for lightweight expansion of MASNUM on a single multi-GPU node APPLE-MASNUM：在单个多gpu节点上加速MASNUM轻量级扩展的并行处理

IF 3.1 3区地球科学 Q2 METEOROLOGY & ATMOSPHERIC SCIENCES

Ocean Modelling

Pub Date : 2025-07-01 Epub Date: 2025-05-02 DOI: 10.1016/j.ocemod.2025.102557

Qi Lou , Changmao Wu , Changming Dong , Xingru Feng , Yuanyuan Xia , Li Liu , Zhengwei Xu , Xu Gao , Meng Sun , Xunqiang Yin

The Marine Science and Numerical Modeling (MASNUM) system, developed for oceanic wave forecasting, play an important role in marine disaster prevention and maritime activities. However, its application is hampered by the requirement of large computing resources. To overcome these barriers, we have implemented an accelerating parallel processing for lightweight expansion of MASNUM (APPLE-MASNUM) on a single compute node with multiple GPUs. In initiating our approach, the mathematical-physics equations of the MASNUM system are thoroughly analyzed to pinpoint the primary computational bottlenecks. This study then transforms MASNUM from a multi-process MPI program into a preliminary GPU-compatible algorithms. Subsequently, the paper proposes an optimization strategy for two-dimensional four-point stencil computations. Following this, an optimization method for overlapping computation with communication is introduced. Finally, a refined data layout scheme tailored for GPUs is designed and implemented. Three numerical experiments with five-day wave forecasts demonstrated that compared to single-core MASNUM, the acceleration ratios of the framework presented in this study are 49.29-fold, 62.58-fold, and 65.74-fold, respectively. This considerable performance boost highlights the efficiency of the lightweight APPLE-MASNUM framework introduced in this research. This signifies the first implementation and optimization of the MASNUM model on a GPU-based heterogeneous platform.

海洋科学与数值模拟（MASNUM）系统是为海浪预报而开发的，在海洋灾害预防和海洋活动中发挥着重要作用。然而，它的应用受到大量计算资源需求的阻碍。为了克服这些障碍，我们在具有多个gpu的单个计算节点上实现了MASNUM （APPLE-MASNUM）轻量级扩展的加速并行处理。在启动我们的方法时，对MASNUM系统的数学物理方程进行了彻底的分析，以确定主要的计算瓶颈。然后，本研究将MASNUM从一个多进程MPI程序转换为一个初步的gpu兼容算法。随后，提出了一种二维四点模板计算的优化策略。在此基础上，提出了一种带通信的重叠计算优化方法。最后，设计并实现了一种适合gpu的精细数据布局方案。3个5天波预报的数值实验表明，与单核MASNUM相比，本文提出的框架的加速比分别为49.29倍、62.58倍和65.74倍。这种相当大的性能提升突出了本研究中引入的轻量级APPLE-MASNUM框架的效率。这标志着MASNUM模型首次在基于gpu的异构平台上实现和优化。

{"title":"APPLE-MASNUM: Accelerating parallel processing for lightweight expansion of MASNUM on a single multi-GPU node","authors":"Qi Lou , Changmao Wu , Changming Dong , Xingru Feng , Yuanyuan Xia , Li Liu , Zhengwei Xu , Xu Gao , Meng Sun , Xunqiang Yin","doi":"10.1016/j.ocemod.2025.102557","DOIUrl":"10.1016/j.ocemod.2025.102557","url":null,"abstract":"<div><div>The Marine Science and Numerical Modeling (MASNUM) system, developed for oceanic wave forecasting, play an important role in marine disaster prevention and maritime activities. However, its application is hampered by the requirement of large computing resources. To overcome these barriers, we have implemented an accelerating parallel processing for lightweight expansion of MASNUM (APPLE-MASNUM) on a single compute node with multiple GPUs. In initiating our approach, the mathematical-physics equations of the MASNUM system are thoroughly analyzed to pinpoint the primary computational bottlenecks. This study then transforms MASNUM from a multi-process MPI program into a preliminary GPU-compatible algorithms. Subsequently, the paper proposes an optimization strategy for two-dimensional four-point stencil computations. Following this, an optimization method for overlapping computation with communication is introduced. Finally, a refined data layout scheme tailored for GPUs is designed and implemented. Three numerical experiments with five-day wave forecasts demonstrated that compared to single-core MASNUM, the acceleration ratios of the framework presented in this study are 49.29-fold, 62.58-fold, and 65.74-fold, respectively. This considerable performance boost highlights the efficiency of the lightweight APPLE-MASNUM framework introduced in this research. This signifies the first implementation and optimization of the MASNUM model on a GPU-based heterogeneous platform.</div></div>","PeriodicalId":19457,"journal":{"name":"Ocean Modelling","volume":"196 ","pages":"Article 102557"},"PeriodicalIF":3.1,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143911537","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Attention-enhanced deep learning model for reconstruction and downscaling of thermocline depth in the tropical Indian Ocean 热带印度洋温跃层深度重建和降尺度的注意力增强深度学习模型

IF 3.1 3区地球科学 Q2 METEOROLOGY & ATMOSPHERIC SCIENCES

Ocean Modelling

Pub Date : 2025-07-01 Epub Date: 2025-03-13 DOI: 10.1016/j.ocemod.2025.102537

Zhongkun Feng , Jifeng Qi , Delei Li , Bowen Xie , Guimin Sun , Baoshu Yin , Shuguo Yang

Accurate estimation of high-resolution thermocline depth is important for investigating ocean processes and climate variability on multiple scales. Due to the sparse coverage and high costs associated with in situ observations, reconstructing ocean interior structure from sea surface data serves as a valuable alternative. In this study, a new deep learning model named Enhanced Block Attention Module-Convolutional Neural Network (EBAM-CNN) was proposed to reconstruct thermocline depth in the tropical Indian Ocean (TIO) from 1993 to 2022. Absolute dynamic topography (ADT), sea surface temperature (SST), and sea surface wind (SSW), along with geographic information (latitude and longitude) and temporal data, were employed as input variables. In comparison with the traditional convolutional neural network (CNN) model, the proposed model demonstrates better performance, with an overall Root Mean Square Error (RMSE) of 5.29 m and a Pearson Correlation Coefficient (R) of 0.87. In addition, this study employs a downscaling approach to reconstruct higher-resolution thermocline depth data. An analysis of the downscaling results confirmed that the proposed framework effectively reconstructed mesoscale sea subsurface features from high-resolution surface observations, significantly enhancing thermocline depth estimates and providing robust data support for oceanic and climatic research.

高分辨率温跃层深度的准确估计对于在多尺度上研究海洋过程和气候变率具有重要意义。由于现场观测的覆盖范围稀疏且成本高，因此从海面数据重建海洋内部结构是一种有价值的替代方法。本研究提出了一种新的深度学习模型——增强块注意模块-卷积神经网络（EBAM-CNN），用于重建1993 - 2022年热带印度洋（TIO）的温跃层深度。以绝对动力地形（ADT）、海面温度（SST）和海面风（SSW）以及地理信息（经纬度）和时间数据作为输入变量。与传统的卷积神经网络（CNN）模型相比，该模型表现出更好的性能，总体均方根误差（RMSE）为5.29 m， Pearson相关系数(R)为0.87。此外，本研究采用降尺度方法重建更高分辨率的温跃层深度数据。对降尺度结果的分析证实，该框架有效地从高分辨率地面观测中重建了中尺度海洋地下特征，显著提高了温跃层深度估算，为海洋和气候研究提供了有力的数据支持。

{"title":"Attention-enhanced deep learning model for reconstruction and downscaling of thermocline depth in the tropical Indian Ocean","authors":"Zhongkun Feng , Jifeng Qi , Delei Li , Bowen Xie , Guimin Sun , Baoshu Yin , Shuguo Yang","doi":"10.1016/j.ocemod.2025.102537","DOIUrl":"10.1016/j.ocemod.2025.102537","url":null,"abstract":"<div><div>Accurate estimation of high-resolution thermocline depth is important for investigating ocean processes and climate variability on multiple scales. Due to the sparse coverage and high costs associated with in situ observations, reconstructing ocean interior structure from sea surface data serves as a valuable alternative. In this study, a new deep learning model named Enhanced Block Attention Module-Convolutional Neural Network (EBAM-CNN) was proposed to reconstruct thermocline depth in the tropical Indian Ocean (TIO) from 1993 to 2022. Absolute dynamic topography (ADT), sea surface temperature (SST), and sea surface wind (SSW), along with geographic information (latitude and longitude) and temporal data, were employed as input variables. In comparison with the traditional convolutional neural network (CNN) model, the proposed model demonstrates better performance, with an overall Root Mean Square Error (RMSE) of 5.29 m and a Pearson Correlation Coefficient (R) of 0.87. In addition, this study employs a downscaling approach to reconstruct higher-resolution thermocline depth data. An analysis of the downscaling results confirmed that the proposed framework effectively reconstructed mesoscale sea subsurface features from high-resolution surface observations, significantly enhancing thermocline depth estimates and providing robust data support for oceanic and climatic research.</div></div>","PeriodicalId":19457,"journal":{"name":"Ocean Modelling","volume":"196 ","pages":"Article 102537"},"PeriodicalIF":3.1,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143684506","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Surface drifter trajectory prediction in the Gulf of Mexico using neural networks 基于神经网络的墨西哥湾海面漂船轨迹预测

IF 3.1 3区地球科学 Q2 METEOROLOGY & ATMOSPHERIC SCIENCES

Ocean Modelling

Pub Date : 2025-07-01 Epub Date: 2025-04-11 DOI: 10.1016/j.ocemod.2025.102543

Matthew D. Grossi , Stefanie Jegelka , Pierre F.J. Lermusiaux , Tamay M. Özgökmen

Machine learning techniques are applied to Lagrangian trajectory reconstructions, which are important in oceanography for providing guidance to search and rescue efforts, forecasting the spread of harmful algal blooms, and tracking pollutants and marine debris. This study evaluates the ability of two types of neural networks for learning ocean trajectories from nearly 250 surface drifters released during the Grand Lagrangian Deployment in the Gulf of Mexico from Jul-Oct 2012. First, simple fully connected neural networks were trained to predict an individual drifter’s trajectory over 24 h and 5 d time windows using only that drifter’s previous velocity time series. These networks, despite having successfully learned modeled trajectories in a previous study, failed to outperform common autoregressive models in any of the tests conducted. This was true even when drifters were pre-sorted into geospatial groups based on past trajectories and different networks were trained on each group to reduce the variability that each network had to learn. In contrast, a more sophisticated social spatio-temporal graph convolutional neural network (STN), originally developed for learning pedestrian trajectories, demonstrated greater potential due to two important features: learning spatial and temporal patterns simultaneously, and sharing information between similarly-behaving drifters to facilitate the prediction of any particular drifter. Position prediction errors averaged around 60 km at day 5, roughly 20 km lower than autoregression, and even better for certain subsets of drifters. The passage of Tropical Cyclone Isaac over the drifter array as a tropical storm and Category 1 hurricane provided a unique opportunity to also explore whether these models would benefit from adding wind as a predictor when making short 24 h predictions. The STNs were found to not benefit from wind on average, though certain subsets of drifters exhibited slightly lower reconstruction errors at hour 24 with the addition of wind.

机器学习技术应用于拉格朗日轨迹重建，这在海洋学中非常重要，可以为搜索和救援工作提供指导，预测有害藻华的扩散，以及跟踪污染物和海洋垃圾。本研究评估了两种类型的神经网络学习海洋轨迹的能力，这些海洋轨迹来自2012年7月至10月墨西哥湾大拉格朗日部署期间释放的近250个海面漂浮物。首先，简单的全连接神经网络被训练来预测单个漂浮物在24小时和5天时间窗口内的轨迹，仅使用该漂浮物之前的速度时间序列。尽管这些网络在之前的研究中成功地学习了建模轨迹，但在进行的任何测试中，它们的表现都不及普通的自回归模型。即使根据过去的轨迹将漂流者预先分类为地理空间组，并对每个组进行不同的网络训练，以减少每个网络必须学习的可变性，情况也是如此。相比之下，一种更复杂的社会时空图卷积神经网络（STN），最初是为学习行人轨迹而开发的，由于两个重要特征，显示出更大的潜力：同时学习空间和时间模式，以及在行为相似的漂移者之间共享信息，以促进对任何特定漂移者的预测。位置预测误差在第5天平均约为60公里，比自回归低约20公里，对于某些漂移子集甚至更好。热带气旋艾萨克（Isaac）以热带风暴和一级飓风的形式经过漂移阵列，为探索这些模型在进行短24小时预测时是否会从增加风作为预测因子中受益提供了一个独特的机会。发现stn平均不受益于风，尽管某些漂移亚群在24小时有风的情况下表现出稍低的重建误差。

{"title":"Surface drifter trajectory prediction in the Gulf of Mexico using neural networks","authors":"Matthew D. Grossi , Stefanie Jegelka , Pierre F.J. Lermusiaux , Tamay M. Özgökmen","doi":"10.1016/j.ocemod.2025.102543","DOIUrl":"10.1016/j.ocemod.2025.102543","url":null,"abstract":"<div><div>Machine learning techniques are applied to Lagrangian trajectory reconstructions, which are important in oceanography for providing guidance to search and rescue efforts, forecasting the spread of harmful algal blooms, and tracking pollutants and marine debris. This study evaluates the ability of two types of neural networks for learning ocean trajectories from nearly 250 surface drifters released during the Grand Lagrangian Deployment in the Gulf of Mexico from Jul-Oct 2012. First, simple fully connected neural networks were trained to predict an individual drifter’s trajectory over 24h and 5d time windows using only that drifter’s previous velocity time series. These networks, despite having successfully learned modeled trajectories in a previous study, failed to outperform common autoregressive models in any of the tests conducted. This was true even when drifters were pre-sorted into geospatial groups based on past trajectories and different networks were trained on each group to reduce the variability that each network had to learn. In contrast, a more sophisticated social spatio-temporal graph convolutional neural network (STN), originally developed for learning pedestrian trajectories, demonstrated greater potential due to two important features: learning spatial and temporal patterns simultaneously, and sharing information between similarly-behaving drifters to facilitate the prediction of any particular drifter. Position prediction errors averaged around 60km at day 5, roughly 20km lower than autoregression, and even better for certain subsets of drifters. The passage of Tropical Cyclone Isaac over the drifter array as a tropical storm and Category 1 hurricane provided a unique opportunity to also explore whether these models would benefit from adding wind as a predictor when making short 24h predictions. The STNs were found to not benefit from wind on average, though certain subsets of drifters exhibited slightly lower reconstruction errors at hour 24 with the addition of wind.</div></div>","PeriodicalId":19457,"journal":{"name":"Ocean Modelling","volume":"196 ","pages":"Article 102543"},"PeriodicalIF":3.1,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143848348","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0