PCMDI Metrics Package

PCMDI Metrics Package
Name	PCMDI Metrics Package
Developer	Program for Climate Model Diagnosis and Intercomparison
Released	200x
Latest release	varies
Programming language	Python, Fortran, C
Operating system	Cross-platform
License	Open source

PCMDI Metrics Package The PCMDI Metrics Package is a software toolkit for evaluating climate model output developed by the Program for Climate Model Diagnosis and Intercomparison as part of efforts tied to the Coupled Model Intercomparison Project and the Intergovernmental Panel on Climate Change. It provides standardized metrics used by researchers affiliated with institutions such as Lawrence Livermore National Laboratory, National Center for Atmospheric Research, University Corporation for Atmospheric Research, and agencies including the National Aeronautics and Space Administration and the National Oceanic and Atmospheric Administration. The package supports analyses relevant to assessments by the World Meteorological Organization, the European Centre for Medium-Range Weather Forecasts, and regional groups like the Met Office.

Overview

The package automates quantitative evaluation of climate simulations against observations from sources like the Global Climate Observing System, National Centers for Environmental Information, and satellite missions such as Landsat, Aqua (satellite), and Terra (satellite). It implements comparison workflows used in reports by the Intergovernmental Panel on Climate Change and benchmarks employed by the Coupled Model Intercomparison Project Phase 5 and Coupled Model Intercomparison Project Phase 6. Stakeholders include researchers at Scripps Institution of Oceanography, Princeton University, Massachusetts Institute of Technology, and policy advisors at the United Nations Framework Convention on Climate Change.

Features and Capabilities

The toolkit provides metrics for fields including surface temperature, precipitation, and atmospheric circulation, drawing observational references from groups like the Global Precipitation Climatology Project, European Space Agency, and NOAA Climate Prediction Center. It offers statistical measures such as bias, root-mean-square error, pattern correlation, and Taylor diagrams used in studies at Stanford University, Columbia University, and Imperial College London. Output formats integrate with visualization tools produced by teams at Matplotlib development team, Panoply, and the Python Software Foundation ecosystem, enabling reproducible diagnostics for consortia such as the World Climate Research Programme.

Architecture and Implementation

Built primarily in Python (programming language), with components in Fortran and C (programming language), the package leverages libraries maintained by the NumPy developers, SciPy community, and xarray developers for array operations and data handling. It interfaces with data conventions defined by the Climate and Forecast (CF) metadata conventions and file formats used by the Network Common Data Form as curated by groups like the Unidata Program Center. Workflows often execute on high-performance computing resources at centers such as Oak Ridge National Laboratory, Argonne National Laboratory, and NERSC.

Use Cases and Applications

Researchers deploy the package to compare historical simulations from model intercomparison projects against reanalysis datasets produced by ERA-Interim, MERRA-2, and JRA-55. Climate centers use it to evaluate model upgrades in workflows adopted by Hadley Centre, Canadian Centre for Climate Modelling and Analysis, and Max Planck Institute for Meteorology. The metrics inform impact assessments used by entities like the World Bank, European Environment Agency, and regional climate services coordinated through the Climate Services Partnership.

Development and Governance

Development is coordinated within the Program for Climate Model Diagnosis and Intercomparison with contributions from researchers at Lawrence Berkeley National Laboratory, University of California, Los Angeles, and University of Washington. Governance follows community-driven practices similar to those of the Open Science Grid and engages with standardization efforts led by the World Meteorological Organization and the Global Climate Observing System. Version control and collaborative development use platforms popularized by the Linux Foundation and the Apache Software Foundation model of open source stewardship.

Adoption and Impact

The package underpins model evaluation in assessments by the Intergovernmental Panel on Climate Change and supports model intercomparison activities tied to the Coupled Model Intercomparison Project Phase 6 and regional downscaling projects coordinated by CORDEX. Its metrics have been cited in studies from institutions such as University of Oxford, Yale University, and Brown University, influencing model development at laboratories like Los Alamos National Laboratory and informing decision-making for organizations including the International Monetary Fund and the United Nations Environment Programme.

Category:Climate software Category:Earth sciences software Category:Open-source software