Integrating Recurrent Neural Networks with Data Assimilation for Scalable Data-Driven State Estimation
This provides a scalable method for state estimation in numerical weather prediction, which is domain-specific but potentially impactful for weather forecasting applications.
The paper tackles the problem of state estimation in numerical weather prediction by integrating recurrent neural networks with data assimilation to replace traditional numerical models, achieving scalable data-driven state estimation that can initialize short-term forecasts without conventional forecast models.
Data assimilation (DA) is integrated with machine learning in order to perform entirely data-driven online state estimation. To achieve this, recurrent neural networks (RNNs) are implemented as surrogate models to replace key components of the DA cycle in numerical weather prediction (NWP), including the conventional numerical forecast model, the forecast error covariance matrix, and the tangent linear and adjoint models. It is shown how these RNNs can be initialized using DA methods to directly update the hidden/reservoir state with observations of the target system. The results indicate that these techniques can be applied to estimate the state of a system for the repeated initialization of short-term forecasts, even in the absence of a traditional numerical forecast model. Further, it is demonstrated how these integrated RNN-DA methods can scale to higher dimensions by applying domain localization and parallelization, providing a path for practical applications in NWP.