The DTC is pleased to announce the release of the multi-component verification framework called the enhanced Model Evaluation Tools (MET), or METplus (https://doi.org/10.5281/zenodo.5567804). METplus contains a suite of Python wrappers and ancillary scripts to enhance the user's ability to quickly set-up and run MET. METplus also has an analysis suite including METviewer and METexpress user interfaces and METdataio, METcalcpy, and METplotpy as shared packages for loading and storing MET output as well as aggregating and plotting results. This coordinated release includes:
- METplus version 5.0 - Release Notes
- MET version 11.0 - Release Notes
METplus Analysis Suite, comprised of:
- METviewer version 5.0 - Release Notes
- METexpress version 5.0 - Release Notes
- METdataio 2.0 - Release Notes
- METcalcpy 2.0 - Release Notes
- METplotpy 2.0 - Release Notes
Highlights –
METplus Wrappers
- Updated Framework:
- NOTE: Use cases that include EnsembleStat wrapper will require config file updates. See METplus Wrappers Upgrade Instructions
- Additional changes: The LOOP_ORDER config variable was removed. This may change the order that commands are executed in a use case, but it should not change the results., METplus Dockerfile was moved to internal/scripts/docker,, default value of SCRUB_STAGING_DIR is now True, default value of METPLUS_CONF now includes the LOG_TIMESTAMP so each METplus run will generate a unique final config file, e.g. metplus_final.conf.20220921121733.
- Updated Wrappers: Enhance MODE wrapper to support multi-variate MODE, allow FCST_IS_PROB variable setting specific to tool (FCST_<tool_name>_IS_PROB), enhance climatology field settings to be consistent with fcst/obs field, update the EnsembleStat wrapper and use case examples to remove ensemble post processing logic
- New use cases: PANDA-C, MJO-ENSO diagnostics, Probability of Exceedence for 85th percentile temperatures, FV3 Physics Tendency plotting, Python Embedding using IODA v2.0, Python Embedding to read native Met Office LFRic grid (u-grid),
MET
- Updated Capability: Restructure the contents of the MET repository so that it matches the existing release tarfiles, remove ensemble post-processing from the Ensemble-Stat tool, enhanced MET’s NetCDF library interface to support level strings that include coordinate variable values instead of just indexes, added a Python helper script/function to transform point_data objects to met_point_data objects for Python Embedding
- New Statistics: Anomaly Correlation Coefficient to VCNT Line Type, 2x2 HSS calculations to include user-defined EC values, Fair CRPS statistic to the ECNT line type, MAE to the ECNT line type from Ensemble-Stat and for HiRA, Mean Absolute Difference (SPREAD_MD) to the ECNT line type, Bias ratio statistic to the ECNT line type from Ensemble-Stat and for HiRA, SEEPS in Point-Stat and Grid-Stat
- New Support: IODA2NC to support IODA v2.0 format, EPA AirNow ASCII data in ASCII2NC, NDBC buoy data in ASCII2NC
- New Diagnostics: TC-Pairs to read hurricane model diagnostic files (e.g. SHIPS) and TC-Stat to filter the new data, TC-Pairs consensus logic to compute the spread of the location, wind speed, and pressure
METplus Analysis Suite
- Added new line types (related to VCNT, SPREAD, SEEPS, CRPS) to METviewer plotting
- Migrated METexpress to using METcalcpy for calculations
- Create a MET data reformatter to allow METcalcpy and METplotpy to be run on the command line without calls to a database
- Added Zonal and Meridional Mean calculations, a MJO ENSO use case, Revision Series to METviewer, Python-based scorecard capability
Visit the DTC METplus User Support page to find online tutorials, video trainings, and documentation for all components of METplus.
Support for US-based HPCs: Over the next week, METplus 5.0 and components will be installed on several US-based high performance computing platforms (e.g. Cheyenne, Jet, Hera, Acorn, and Frontera) to make it more accessible to both the research and operations community. Please see the list of existing builds. Additionally, an Amazon Web Service (AWS) Machine Image (AMI) will be made available for use with data hosted on AWS. NOTE: METviewer and METexpress are intended to be installed on your local server.
Save these dates: METplus Advanced Training Series begins in February 2023 and will run through May 2023. Sign up for updates on the METplus Website
