Ocean Observation & Modeling Group, North Carolina State University
OceanNet 2.0 is a deep-learning based regional ocean emulator for the Gulf of Mexico/America. It offers long-term ensemble forecasts of daily sea level and ocean current velocity. The forecast system consists of 40 ensemble members initialized from CMEMS global ocean analysis/forecast product. Each ensemble member integrates individually and produces daily forecasts for the next 120 days.
OceanNet 2.0 utilizes advanced neural network architectures to emulate ocean dynamics with high accuracy and computational efficiency compared to traditional numerical models.
[1] Chattopadhyay, A., M. Gray, T. Wu, A. B. Lowe, R. He. (2024). OceanNet: A principled neural operator-based digital twin for regional oceans, Scientific Reports, 14, 21181 (2024) doi: 10.1038/s41598-024-72145-0
[2] Lowe, A. B., M, Gray, A. Chattopadhyay, T. Wu, R. He. (2025). Long-term predictions of Loop Current Eddy evolutions using OceanNet: a Fourier neural operator-based data driven ocean emulator, Artificial Intelligence for the Earth System, in press. https://doi.org/10.1175/AIES-D-24-0039.1
[3] Gray, M. A., A. Chattopadhyay, T. Wu, A. Lowe, and R. He. (2025). Long-term Prediction of the Gulf Stream Meander Using OceanNet: a Principled Neural Operator-based Digital Twin, Ocean Sciences, in press. https://doi.org/10.5194/egusphere-2024-1238
[4] L.-J., Leonard, M. Darman, S. Hazarika, T. Wu, M. Gray, R. He, A. Wong, and A. Chattopadhyay. (2025). Simultaneous emulation and downscaling with physically-consistent deep learning-based regional ocean emulators. arXiv preprint. https://arxiv.org/abs/2501.05058
• Accurate 120-day long-term forecasts
• Ensemble uncertainty quantification
• Daily sea surface height and velocity forecasts
• Specialized in predicting ocean dynamics in the Gulf
Ensemble means of SSH and probable location of Loop Current (LC) and Loop Current Eddy (LCE) based on SSH threshold. Shows probability forecasting with color shading and contour analysis.
Ensemble means and spread (standard deviation) of SSH and sea surface speed. Provides uncertainty quantification for model predictions.
Time evolution of 120-day SSH forecast from all 40 ensemble members. Focused on Loop Current System dynamics and variability.
OOMG (Ocean Observation & Modeling Group) develops state-of-the-art ocean models including ensemble forecasting and data assimilation systems. Our models integrate physical oceanography, machine learning, and numerical methods to deliver accurate ocean predictions.
OceanNet utilizes ensemble predictions with multiple model runs for uncertainty quantification, while CNAPS2 focuses on data assimilation techniques that incorporate real-time observations to improve forecast accuracy and provide comprehensive ocean analysis.
Both OOMG models provide global ocean predictions including current patterns, temperature distributions, sea surface height, and other key oceanographic variables with high spatial and temporal resolution for comprehensive ocean research and applications.
OOMG model predictions support marine operations, climate research, fisheries management, and environmental monitoring. Our research helps scientists and decision-makers understand ocean dynamics and their impacts on weather, climate, and marine ecosystems.