ACCESS-ENSO-Recipes: A Flexible Workflow for ENSO and IOD Evaluation Using ESMValTool and Jupyter Notebooks

The El Niño-Southern Oscillation (ENSO) is the dominant mode of interannual variability, with global climate impacts that underscore the importance of its accurate representation in climate models. Building on the success of the CLIVAR ENSO metrics package, our work focuses on developing an advanced workflow for evaluating ENSO in the ACCESS family of models, extending its functionality to include diagnostics for the Indian Ocean Dipole (IOD).

Our approach leverages the IRIS-based pre-processors within ESMValTool, enabling a modular and flexible development of ENSO and IOD diagnostics. These are implemented as Python-based diagnostics that can seamlessly integrate into both exploratory Jupyter Notebooks and the traditional YAML-based ESMValTool recipes. The notebooks, part of the open-source ACCESS-ENSO-Recipes, are hosted on GitHub and serve as a powerful platform for designing diagnostics, visualisation, and interactive data exploration. Meanwhile, the YAML recipes facilitate the semi-automated evaluation of multi-model ensembles and large datasets, ensuring compatibility with established workflows for climate model evaluation.

Our current suite of diagnostics aims to reproduce the functionality of the CLIVAR ENSO metrics package, focusing on ENSO variability, teleconnections, and physical processes. By utilising notebooks, we create an agile environment for developing and refining diagnostic tools, enhancing collaboration between scientists and model developers. At the same time, the structured recipe format ensures reproducibility and scalability, enabling systematic analysis across models and ensembles.

We plan to extend this approach to evaluate broader oceanic and atmospheric processes in the ACCESS models, enabling more comprehensive assessments of model performance. The ACCESS-ENSO-Recipes and Python diagnostics will also be shared on the ESMValTool GitHub repository to encourage collaboration and wider use in the climate modelling community.

This workflow balances innovation and scalability by integrating flexible notebooks with structured legacy tools. Advancing ENSO and IOD evaluation will improve climate model accuracy and our understanding of their impacts on current and future climates.