tenomoto/l63vda

This repository contains the Python version of the Lorenz-63 model in Huang and Yang (1996). The code was used in the 25-th data assimilation summer school in 2021 held online in Japan.

Jupyter notebooks

Note that initial condition files need to be created manually in Binder.

• Variational data assimilation

• original version generating files: nbviewer
• modified version without files: nbviewer

• Ensemble Kalman filter with perturbed observations nbviewer
• Neural network nbviewer
• Autmatic differentiation with JAX nbviewer

References

• Huang, X.-Y. and X. Yang, 1996: Variational data assimilation with the Lorenz model. Technical Report 26, HIRAM, April 1996, 44 pp.