Momentum Training - LFRic Atmosphere

Momentum^® is a software framework for modelling Earth’s environment, developed and used by the Momentum Partnership. The framework includes rigorously evaluated Science Configurations, which define how to configure components of the framework to build prediction and projection systems, both regional and global.

Momentum follows a seamless modelling approach, using the same model components across temporal and spatial scales, similar to the Unified Model. The framework includes model components for the atmosphere, land surface, ocean, sea-ice, and other parts of the Earth system. It contains software for data assimilation, verification, and technical tasks like workflow management.

LFRic Atmosphere is the atmospheric model component of Momentum. It is the successor to the Unified Model. See the Momentum website for information about the other components of the modelling framework.

Who this course is for

This course is designed for new users of LFRic Atmosphere who have a general background in weather or climate science. You should be comfortable working in a Linux terminal and have some experience of Python.

The course also uses Cylc, Rose, version control (FCM and Git) and Jupyter notebooks. You do not need to know these in advance. Where a section relies on one of them, we explain what you need and suggest where to find out more. Links to introductory material are listed under Further reading at the end of this page.

Tip

For help, questions, or feedback on the training, please use GitHub Discussions. You are welcome to start a discussion without opening an issue or preparing a pull request.

Choosing your platform

Many of the practical pages give instructions that depend on the computer you are working on. Those instructions are presented in tabs. Pick the tab that matches your platform once, and the rest of the site follows your choice.

Tab	Use it when
Met Office	You are working on a Met Office system, using the managed LFRic environment and the Met Office module stack.
Monsoon	You are working on Monsoon3, the collaborative HPC service shared with Momentum partners. Monsoon3 splits access, interactive workflow management, and HPC work across separate services, so some commands must be run from a particular service. See Using Monsoon3.
Other	You are working on your own machine, or on a system run by another institution. You provide the environment yourself.

Important

Support for the Other route is limited. It is enough for the visualisation and analysis material, where you only need Python, Iris, and the training repository. It is not enough for the exercises that build and run the model: those need a supported LFRic build environment, Met Office repository access, shared ancillary data, and HPC batch queues. Where an exercise cannot be completed on your own machine, the Other tab says so.

Contents of the training course

• 1. Introduction to LFRic Atmosphere
  • Seamless Modelling
  • Components, Science Configurations, and Systems
  • Land Surface and JULES
  • Ocean and Sea Ice
  • Rivers
  • LFRic Atmosphere features and history
  • Introduction review quiz
• 2. Unstructured Grid (UGRID) - LFRic output format
  • Structured grid
  • Unstructured mesh
  • Practical using unstructured data
• 3. LFRic code and software
  • Platform and prerequisites
• 4. Science Configurations
  • Introduction to Science Configurations
  • Tools for Modelling
  • Evaluation and Development
  • Exercises in Global Configurations
  • Exercises in Regional Configurations
  • Exercises in Idealised Configurations
• Glossary
• Appendices
  • Using Monsoon3
  • Using Jupyter notebooks
• Publication References

Further reading

See also

Introductory material for tools used in the course:

• Cylc Tutorial Cylc is a workflow engine that runs tasks according to schedules and dependencies.

• Rose Tutorial Rose is a toolkit for writing, editing, and running application configurations.

• FCM user guide and Git novice lesson: tools for tracking and managing changes in code.

• Using Jupyter notebooks: how to launch JupyterLab, choose the correct kernel, run notebook cells, and find beginner-friendly Python resources.

Background on Momentum:

• Momentum Partnership

• Unified Model