Python program for EOF analysis of MIKE 21 data to enable training of Sparse Gaussian Process (SPGP / Sparse GP) models to predict flood inundation extent.
The model is developed for the publication: "Upskilling low-fidelity hydrodynamic models of flood inundation through spatial analysis and Gaussian Process learning" and uses the Chowilla floodplain as case study.
MIKE 21 result files needed to run the code are available at https://doi.org/10.26188/62143bdd9fa63.