ESMValCore and ESMValTool: analyzing CMIP data made easy

Abstract:

The Earth System Model Evaluation Tool (ESMValTool) is a free and open-source community diagnostic and performance metrics tool for the evaluation of Earth system models participating in the Coupled Model Intercomparison Project (CMIP). Version 2 of the tool (www.esmvaltool.org) features a brand new design, consisting of ESMValCore (https://github.com/esmvalgroup/esmvalcore), a package for working with CMIP data and ESMValTool (https://github.com/esmvalgroup/esmvaltool), a package containing the scientific analysis scripts. This new version has been specifically developed to handle the increased data volume of CMIP Phase 6 (CMIP6) and the related challenges posed by the analysis and the evaluation of output from multiple high-resolution or complex Earth system models. The tool also supports CMIP5 and CMIP3 datasets, as well as a large number of re-analysis and observational datasets that can be formatted according to the same standards (CMOR) on-the-fly or through scripts currently included in the ESMValTool package.

At the heart of this new version is the ESMValCore software package, which provides a user configurable framework for finding CMIP files using a “data reference syntax”, applying commonly used pre-processing functions to them, running analysis scripts, and recording provenance. Numerous pre-processing functions, e.g. for data selection, regridding, and basic statistics are readily available and the modular design makes it easy to add more. The ESMValCore package is easy to install with relatively few dependencies, written in Python 3 and based on state-of-the-art open-source libraries. The diagnostic part of the ESMValTool includes a large collection of standard recipes for reproducing peer-reviewed analyses for many quantities and variables across atmosphere, ocean, and land domains, diagnostics and performance metrics focusing on the mean-state, trends, variability and important processes, phenomena, as well as emergent constraints. These diagnostics are partly based on model-to-model comparisons, but many also use observational data sets such as satellite and ground-based observations for model evaluation. Documentation of both the ESMValCore and ESMValTool is available at https://docs.esmvaltool.org.