Tip
Useful links: Documentation | Source Code | Issue Tracker | Releases | Discussion Forum
The Community Seamless Evaluation Toolkit, CSET, is a community-developed open source toolkit for evaluation, verification, and investigation of weather and climate models. It supports the evaluation of physical numerical models, machine learning models, and observations seamlessly across time and space scales. CSET primarily targets, but is not limited to, high-resolution atmospheric processes, from convective to turbulence scales (i.e. kilometre to sub-kilometre grid spacing), across regional or global domains.
CSET provides a centralised and peer-reviewed set of tools to aid process-oriented verification and evaluation for UM, LFRic, and machine learning models, supporting both deterministic and ensemble configurations.
At the Met Office and Momentum® Partnership CSET supports parametrisation development, diagnostic development and evaluation research. It is integral to the Regional Atmosphere and Land (RAL) model development process for the Unified Model and LFRic atmospheric modelling codes.
CSET is designed to be continuously evolving and improving, driven by community inputs. Support for verification and evaluation of a range of machine learning models is expected to grow, alongside use of observations from an increasing range of sources to support evaluation. It will utilise the Model Evaluation Tools (MET) software to provide a range of verification metrics aligned with operational verification best practices. Where relevant, CSET will provide interfaces to utilise other evaluation packages to support particular evaluation requirements.
Please visit the documentation to learn more about CSET and how to use it.
If you want to ask or share with the CSET community, please use the relevant category of the Simulation Systems Discussions Forum.
Evaluation software developed for and by a wide network of model development and evaluation scientists, enabling common approaches to distributed evaluation activities.
Supporting assessment, evaluation, verification and understanding of physical and machine learning models as well as observations across time and space scales, and from regional to global application.
Providing a process-oriented focus to model assessment, supporting depth of comparison between different model configurations and assessment relative to a range of observations.
A flexible software including code, recipes, diagnostics and workflow to manage a range of user requirements, underpinned by modern software development practices.
Contributions are readily welcomed! Visit the contributing documentation to get started with developing CSET.
In addition to reading the working practices, the key recommendation is early communication. Open an issue on GitHub while your proposed change or addition in the design phase so others can provide guidance.
© Crown copyright, Met Office (2022-2025) and CSET contributors.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.
GitHub Copilot was used in the development of this software.