Objectively summarizing Earth System Model (ESM) behavior has become increasingly challenging with the rapid increase in the number of experiments, enhanced complexity and increasing resolution. The next phase of the Coupled Model Intercomparison Project (CMIP6) will soon provide an unprecedented amount of data to the research community, with 287 experiments, from 100 models, and data volume estimates projecting as much as 50 petabytes.

In this presentation we highlight the PCMDI Metrics Package (PMP) which is designed to produced a diverse suite of relatively robust objective comparisons between CMIP class models and the latest available observations. The PMP exploits DOE supported python based tools that have been optimized to work with the CMIP data structures and metadata conventions. The PMP statistical summaries include a broad range of earth system characteristics (spanning the atmosphere, land, ocean and sea-ice), across a broad range of space and time scales.

These summaries provide objective model evaluation information, mining the CMIP archive to distil petabytes of data into kilobytes of radically condensed but invaluable information about model behavior and performance. The PMP simulation summaries are being prepared with a special emphasis on start-to-finish workflow provenance to facilitate reproducibility, and are being made available to researchers via access to a database of results, code, and a web-based interface to the underlying diagnostics for broader exploration.