v1.1.0

Refactor the code, made available on PyPI.

Refactor the assim_tools module to use classes for components of the core assimilation algorithm.

• Available Assimilators: ETKF, TopazDEnKF, EAKF

• Available Updators: Additive and Alignment

• Misc. transform funcs: null, scale_bandpass

Added readthedocs documentation pages.

Added a few examples of use cases: QG model benchmarking of filter performance

v1.0-beta

v1.0 beta test release

Most of the essential functionality is implemented: serial and batch assimilation algorithms, parallel scheme, model interfaces, dataset interfaces, synthetic observation option, localization. More features will be implemented soon.

See "Quick Start Guide" for instructions to setup the system.

A simple 2d vortex model (models/vort2d) provides a testcase to play with the system. The script "models/vort2d/run_forecast.sh truth" can help perform a deterministic forecast and save it as "truth", then "scripts/run_cycle.sh" performs cycling data assimilation using synthetic observations. The whole procedure is also illustrated step-by-step in a jupyter-notebook in "tutorials/vort2d_testcase.ipynb"

All feedback are welcomed. Please use the Issues tab if you want to contribute to NEDAS developement!