NERC/NCEO/DARC Training course on data assimilation and its interface with machine learning 2026

Practical material for NERC/NCEO/DARC Training course on data assimilation and its interface with machine learning 2026 in Reading

On Google Colab (Recommended)

This year, the DARC training course is splitted into foundation and advanced course. In the foundation course, this repository covers:

• Building blocks of DA: Streamlit and Jupyter Notebook
• Reanalysis: Jupyter Notebook

In the advanced course, we provide practicals for

• Variational DA: Jupyter Notebook and JEDI instructions
• EnKF DA: Jupyter Notebook, Optional Notebook and JEDI instructions
• Machine learning and DA

Running on your own laptop locally

To run practicals on your local laptop, we recommend configure the Python environment using conda. Once you installed conda, you can launch the Jupyter notebook with the following command in your terminal or anaconda command prompt:

# create a new conda envrionment and install required packages
conda create -n da_course conda-forge::numpy conda-forge::scipy conda-forge::matplotlib conda-forge::cartopy conda-forge::ipywidgets conda-forge::jupyter conda-forge::xarray conda-forge::dask conda-forge::netCDF4 conda-forge::bottleneck conda-forge::kagglehub
# activate the conda environment
conda activate da_course
# launch jupyter notebook
jupyter notebook

Setting up connections to Reading Academic Computing Cluster (RACC) -- in-person only

Advanced course participants will obtain a temporary RACC account. To utilise the computing resources on the computing cluster, users must connect to RACC via ssh.

ssh -Y -J USERNAME@arc-ssh.reading.ac.uk USERNAME@racc.rdg.ac.uk