Systematic Bias Correction in Ocean Mesoscale Forecasting Using Machine Learning.

The ocean circulation is modulated by meandering currents and eddies. Forecasting their evolution is a key target of operational models, but their forecast skill remains limited. We propose a machine learning approach that improves the output of an ocean circulation model by learning and predicting its systematic biases. This method can be applied a priori to any region, and is tested in the Gulf of Mexico, where the Loop Current (LC) and the large anticyclonic eddies that detach from it are major forecasting targets. The LC dynamics are recurrent and lie on a low‐dimensional dynamical attractor. Building upon the information gained analyzing this low dimensional attractor, we improve the representation of sea surface anomalies in model outputs through information from satellite altimeter data using a Sequence‐to‐Sequence model, which is a special class of Recurrent Neural Network. Building upon the HYCOM‐NCODA analysis system, we deliver a correction to the forecast at the observation resolution. For at least 15 days the proposed method learns to forecast the systematic bias in the HYCOM‐NCODA, outperforming persistence, and improving the forecast. This data‐driven approach is fast and can be implemented as an added step to any dynamical hindcasting or forecasting model. It offers an interesting avenue for further developing hybrid modeling tools. In these tools, fundamental physical conservations are preserved through the integration of partial differential equations which obey them. In addition, the method highlights specific deficiencies of the hindcast system that deserve further investigation in the future. Plain Language Summary: Predicting the evolution of ocean circulation using a physically constrained models is critical for weather forecasting and for quantifying nutrient transport and water mass exchanges, but models are far from perfect. We propose a generalized data‐driven method to improve a circulation model performance by identifying and predicting its systematic biases, and we test it in the Gulf of Mexico. We use a recurrent neural network to improve and predict the evolution of sea surface anomalies over 7–15 days using information obtained from satellite altimetry data. The results show that our proposed data‐driven method is fast and accurate and can be easily implemented as an additional step to other dynamic models, providing an interesting avenue for further development of hybrid forecasting tools. Key Points: Machine learning tools are used to improve model predictive skill using observationally based data setsOur method has great potential to extend the current prediction length from about 7 days to 11–15 days without significantly compromising accuracy [ABSTRACT FROM AUTHOR]