DFN is a deep neural network that predicts ocean surface flow fields from operational satellite observations that reduces the mean positioning error after a 7-day forecast compared with the operational model and demonstrates the potential of deep learning for ocean surface flow prediction.
Abstract
The North Pacific Subtropical Gyre (NPSG) is a major accumulation zone for floating plastic debris, resulting from basin-scale convergent ocean circulation. Effective cleanup strategies in this region rely on accurate forecasts of Lagrangian particle drift. Here, we introduce Drift Field Net (DFN), a deep neural network that predicts ocean surface flow fields from operational satellite observations. DFN is trained using a novel two-stage strategy that combines pretraining on simulated data with Lagrangian fine-tuning based on an advection-consistent loss function. This physics-informed optimization directly improves the accuracy of particle trajectory predictions. We evaluate DFN against an operational physics-based forecasting system and demonstrate the potential of deep learning for ocean surface flow prediction. On in situ drifter trajectories, DFN reduces the mean positioning error by 20 km after a 7-day forecast compared with the operational model. Furthermore, Lagrangian fine-tuning with the proposed advection loss further reduces the positioning error by 10 km, highlighting the benefits of incorporating Lagrangian constraints into the training process.
This work frames trajectory prediction as a denoising task and presents Conditional Diffusion models for Trajectories (CoDiT), which adapts the denoising diffusion framework, originally developed for image synthesis, to the problem of trajectory forecasting.
Christian Donner, Shirin Goshtasbpour, Emanuele Dalsasso et al.· Machine Learning: Earth· 0 citations
Simulating the trajectory of surface drifters in the ocean matters for search and rescue, pollution tracking, oil and chemical spill response, and marine risk analysis. Accurate prediction remains difficult, as widely acknowledged in the literature, and also illustrated by the ``Forecasting Floats in Turbulence''challe...
This study proposes a novel deep learning framework that uses a U-Net to generate an initial high-resolution SST estimate, which is subsequently refined using a residual corrective approach, and progressively refines initial U-Net predictions by incorporating dynamically scaled residuals at each step, enabling accurate...
Onkar Jadhav, Tim French, I. Janeković et al.· 1 citation
This study confirms that Deepcov-EnKF overcomes key limitations of traditional EDA frameworks, significantly enhances the accuracy of SST assimilation, and lays the foundation for high-precision marine forecasting systems.
Bao-Xu Li, H. Leng, Jun-Qiang Song et al.· 0 citations
High-resolution measurements of near-surface optical turbulence over the Tibetan Plateau are essential for optical propagation studies, astronomical site characterization, and adaptive optics applications. In this study, a multi-level tethered UAV system was deployed in Ngari Prefecture, China, to obtain synchronous in...
Xiaoyu Hu, Yi-Yung Cheng, Feng-Fu Tan et al.· Remote Sensing· 0 citations
Rapid forecasting of spatiotemporal flood evolution is critical for emergency response. However, traditional hydrodynamic models face computational bottlenecks, while purely data-driven models lack physical constraints, frequently generating topologically implausible overspill in complex terrains. Here we propose CA-DS...
Zheng-Hai Zhou, Jian-Zhong Lu, Meng-Fan Li et al.· IEEE Journal of Selected Top...· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduNov 17, 2026
By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.
Microsoft Research Blog· microsoft.comNov 16, 2026
AI coding agents are increasingly help developers fix bugs and build features largely on their own, but every interaction with their models consumes tokens. In each session, an agent sends instructions, tool descriptions, and session history to its model. Most efforts to reduce this cost focus on individual tasks. Yet developers return to the same […] The post TokenEconomizer: Learning from Experience to Make AI Coding Agents More Efficient appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduNov 13, 2026