LangYa: a large AI model for global ocean forecasting
- 1Key Laboratory of Ocean Observation and Forecasting, Laboratory of Ocean Circulation and Waves, Institute of Oceanology, Chinese Academy of Sciences, Qingdao 266071, China; Qingdao Key Laboratory of AI Oceanography, Qingdao 266000, China
- 2Key Laboratory of Ocean Observation and Forecasting, Laboratory of Ocean Circulation and Waves, Institute of Oceanology, Chinese Academy of Sciences, Qingdao 266071, China. Electronic address: [email protected]
- 3Key Laboratory of Ocean Observation and Forecasting, Laboratory of Ocean Circulation and Waves, Institute of Oceanology, Chinese Academy of Sciences, Qingdao 266071, China. Electronic address: [email protected]
Abstract
Ocean forecasting is crucial for both scientific research and societal benefits. Large artificial intelligence (AI)-based models have recently boosted forecasting efficiency and accuracy. However, it remains challenging to develop a comprehensive AI-driven ocean forecasting system capable of integrating cross-spatiotemporal and atmospheric forcing. This study introduces LangYa, a cross-spatiotemporal and atmospheric forcing ocean forecasting system featuring: (1) a large-language-model-based (LLM-based) time embedding to explicitly represent forecast lead times, (2) an asynchronous cross-iterative random sampling strategy to represent the impacts of atmospheric forcing on ocean processes, (3) an ocean self-attention module to enhance network stability and accelerate training convergence, and (4) an adaptive loss function to capture ocean dynamics in the thermocline, at depths ranging from tens of meters to about 300 m. LangYa is trained on 27 years of global ocean data from the Global Ocean Reanalysis and Simulation, version 12 (GLORYS12). Using reanalysis and observational data, compared to existing open-source AI-based forecasting systems and numerical models, LangYa enables a single model to produce forecasts with lead times of 1 to 7 d (1/12°, daily) and achieves 7 d RMSEs below 0.0736 m/s, 0.0701 m/s, 0.4376 ℃, and 0.1302 psu for global currents, temperature, and salinity respectively. These quantitative results indicate that LangYa provides clear advantages in forecast accuracy, lead-time robustness, and stability for global OSV forecasting, demonstrating its potential for real-time operational deployment.
Related
- OceanForecastBench: A Benchmark Data Set for Data‐Driven Global Ocean Forecasting
- A deep learning global ocean forecasting model with sub-daily and eddy-resolving resolution
- Njord: A Probabilistic Graph Neural Network for Ensemble Ocean Forecasting
- LFS-SEnDA: a global ocean forecast system with an efficient Sliding-Ensemble-based Data Assimilation scheme
- Global patterns in small organic particle loss in the mesopelagic ocean
BibTeX
@article{yang2026langya,
title = {{LangYa: a large AI model for global ocean forecasting}},
author = {Nan Yang and Chong Wang and Zimeng Zhao and Meihua Zhao and Huiling Zheng and Bin Zhang and Fan Wang and Xiaofeng Li},
journal = {中国科学通报：英文版},
year = {2026},
month = aug,
doi = {10.1016/j.scib.2026.08.001},
url = {https://doi.org/10.1016/j.scib.2026.08.001},
note = {CC BY-NC},
}
RIS
TY - JOUR
TI - LangYa: a large AI model for global ocean forecasting
AU - Nan Yang
AU - Chong Wang
AU - Zimeng Zhao
AU - Meihua Zhao
AU - Huiling Zheng
AU - Bin Zhang
AU - Fan Wang
AU - Xiaofeng Li
PY - 2026
DA - 2026/08/02
JO - 中国科学通报：英文版
DO - 10.1016/j.scib.2026.08.001
UR - https://doi.org/10.1016/j.scib.2026.08.001
AB - Ocean forecasting is crucial for both scientific research and societal benefits. Large artificial intelligence (AI)-based models have recently boosted forecasting efficiency and accuracy. However, it remains challenging to develop a comprehensive AI-driven ocean forecasting system capable of integrating cross-spatiotemporal and atmospheric forcing. This study introduces LangYa, a cross-spatiotemporal and atmospheric forcing ocean forecasting system featuring: (1) a large-language-model-based (LLM-based) time embedding to explicitly represent forecast lead times, (2) an asynchronous cross-iterative random sampling strategy to represent the impacts of atmospheric forcing on ocean processes, (3) an ocean self-attention module to enhance network stability and accelerate training convergence, and (4) an adaptive loss function to capture ocean dynamics in the thermocline, at depths ranging from tens of meters to about 300 m. LangYa is trained on 27 years of global ocean data from the Global Ocean Reanalysis and Simulation, version 12 (GLORYS12). Using reanalysis and observational data, compared to existing open-source AI-based forecasting systems and numerical models, LangYa enables a single model to produce forecasts with lead times of 1 to 7 d (1/12°, daily) and achieves 7 d RMSEs below 0.0736 m/s, 0.0701 m/s, 0.4376 ℃, and 0.1302 psu for global currents, temperature, and salinity respectively. These quantitative results indicate that LangYa provides clear advantages in forecast accuracy, lead-time robustness, and stability for global OSV forecasting, demonstrating its potential for real-time operational deployment.
SN - 2095-9273
SN - 2095-9281
ER -