Representing ill-known parts of a numerical model using a machine learning approach
This addresses the challenge of incomplete physical models in Earth System science, offering a method to integrate observed data for improved forecasting, though it is incremental as it builds on existing hybrid modeling approaches.
The paper tackles the problem of unknown or poorly represented processes in Earth System numerical models by proposing a hybrid model that combines a physical-based model with a neural network trained from observations. The result shows that the hybrid model accurately reproduces unknown terms (correlation close to 1), maintains key system properties over long-term simulations, and generalizes to new forcings not seen during training.
In numerical modeling of the Earth System, many processes remain unknown or ill represented (let us quote sub-grid processes, the dependence to unknown latent variables or the non-inclusion of complex dynamics in numerical models) but sometimes can be observed. This paper proposes a methodology to produce a hybrid model combining a physical-based model (forecasting the well-known processes) with a neural-net model trained from observations (forecasting the remaining processes). The approach is applied to a shallow-water model in which the forcing, dissipative and diffusive terms are assumed to be unknown. We show that the hybrid model is able to reproduce with great accuracy the unknown terms (correlation close to 1). For long term simulations it reproduces with no significant difference the mean state, the kinetic energy, the potential energy and the potential vorticity of the system. Lastly it is able to function with new forcings that were not encountered during the training phase of the neural network.