Key points are not available for this paper at this time.
The PCMDI Metrics Package (PMP) is an open-source Python-based framework that enables objective "quick-look" comparisons and benchmarking of Earth System Models (ESMs) against observation-based reference datasets. The PMP, which is focused primarily on atmospheric quantities, has been used for routine and systematic evaluation of thousands of simulations from the Coupled Model Intercomparison Project (CMIP). Ongoing work aims for seamless application of the tool to the next generation of CMIP (CMIP7), with an aspiration to aid modeling groups during their development cycle. The latest version of PMP offers a diverse suite of evaluation capabilities covering large- to global-scale climatology and annual cycle, variability modes such as tropical and extratropical variability modes e.g., ENSO and MJO, regional monsoons, cloud feedback, and high frequency characteristics e.g., extremes. Current work is expanding the scope of PMP to include the evaluation of the following: (1) Quasi-Biennial Oscillation (QBO) and its teleconnection to MJO, (2) atmospheric blocking and rivers leveraging Machine Learning based pattern detection algorithms, and (3) polar and high-latitude regions by implementing sectional sea-ice area metrics. The PMP is also advancing its evaluation capabilities to help evaluate higher resolution simulations such as those from the HighResMIP, cloud-resolving E3SM experiments, and regionally downscaled products. This presentation will highlight the motivation for routine model evaluation, introduce the PMP, share progress on current polar metrics, and discuss future plans and opportunities to connect with ongoing efforts.
Lee et al. (Fri,) studied this question.