Output metadata standards

To facilitate provenance within E3SM experiments and model intercomparisons, the following standards should be followed in all files output by E3SM components.

General

All files should use a self-describing data format that allows metadata both per-variable and per-file (aka global metadata).

E3SM uses the NetCDF format for all output. Other formats may be used for initial output (e.g ADIOS) but must be converted to NetCDF for post-processing and archiving.

History global attributes

History files from E3SM components should all have AT LEAST these global attributes for each file. Other’s may be added as needed by the individual model.

Examples are from EAM and should be adjusted as appropriate.

These attributes are a mix of traditional CESM/CAM global attributes and CMIP global attributes. Many CMIP global attributes can only be added/defined during CMOR-ization. See https://docs.google.com/document/d/1h0r8RZr_f3-8egBMMh7aqLwy3snpD6_MrDz1q8n5XUk/edit

Anything labeled as “from coupler” can get its value from the infodata block in the coupler with a call to  seq_infodata_GetData

Format below: attribute name = (description of value) “example”;

title = (what is in this file) "EAM History file information" ;

source = (common full name for model) "E3SM Atmosphere Model" ;

source_id = (the git hash from coupler)

product = "model-output" ;

realm = (see table in https://github.com/WCRP-CMIP/CMIP6_CVs/blob/master/CMIP6_realm.json )

case = (CIME case id from coupler) ;

username = (username from coupler) ;

hostname = (hostname from coupler) ;

git_version = (the git hash from coupler)

history = (date-time stamp of file creation) "created on 05/02/21 23:58:09" ;

Conventions = "CF-1.7" ;

institution_id = "E3SM-Project" ;

institution = "LLNL (Lawrence Livermore National Laboratory, Livermore, CA 94550, USA); ANL (Argonne National Laboratory, Argonne, IL 60439, USA); BNL (Brookhaven National Laboratory, Upton, NY 11973, USA); LANL (Los Alamos National Laboratory, Los Alamos, NM 87545, USA); LBNL (Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA); ORNL (Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA); PNNL (Pacific Northwest National Laboratory, Richland, WA 99352, USA); SNL (Sandia National Laboratories, Albuquerque, NM 87185, USA). Mailing address: LLNL Climate Program, c/o David C. Bader, Principal Investigator, L-103, 7000 East Avenue, Livermore, CA 94550, USA" ;

contact = "e3sm-data-support@listserv.llnl.gov" ;

History variable attributes

Each variable in a history output file from an E3SM component should have AT LEAST the following attributes if defined for the variable. Others may be added as needed by the model.

The missing value or FillValue should be the CMIP standard 1.e+20.

units = (CF-compliant unit if available. See https://cfconventions.org/Data/cf-standard-names/current/build/cf-standard-name-table.html. NOTE a unitless variable should either have no “units” attribute OR the unit should be set to “1”)

long_name = (A long descriptive name which may, for example, be used for labeling plots. Use CMIP long name if available. See )

standard_name = (CF-compliant standard name if available. Use only if units and sign match CF. See )

cell_methods = “time: mean” (for time-averaged history.)

cell_methods = “time: point” (for instantaneous values. Other possibilities are in

(Do not add CF spatial cell_methods for a variable because CF only defines those for rectangular structured grids.)

Restart global attribute

 

Restart variable attributes