ESGF CMOR Data Converter

Picture9The CMOR Data Converter reads in a multitude of standard data formats, such as hdf3, hdf4, a Grads control file, a Matlab data file or a list of these files, and converts the data into the CMIP5 data format to allow publication on the Earth System Grid Federation (ESGF) using the CMOR2 package API developed by Lawrence Livermore National Laboratory.

The CMOR Data Converter automates the use of the CMOR2 program when preparing observational and analysis data for publication into the Earth System Grid Federation (ESGF) so that it can be used for comparison with climate models.

This software allows quick and efficient preparation of the data. The sharing of this software will reduce the need for each organization to duplicate these efforts. The code, written in python, uses tools from NetCDF, which was developed and is maintained at Unidata, part of the University Corporation for Atmospheric Research (UCAR) Community Programs (UCP). It provides a front end interface, or Application Programming Interface to PCMDI’s CMOR2.  The software was developed to be used within the CISTO Climate Data Services and publicly released under NASA’s Opensource Agreement via NASA’s Open Climate Workbench available through the Apache Software Foundation.

The programs reads the input file, identifies the format, makes the necessary unit and grid conversions, performs time aggregation, converts it to CMIP5, and writes the output data in the required directory structure.  CMOR Data Converter is able to convert the following formats to CMIP5:

From Grads control file (ctl) to CMIP5:

  • Time aggregation will be done automatically based on the time definition variable TDEF.

From NetCDF3 or NetCDF4:

  • Only one file at a time will be processed, and time aggregation will be done depending on the “time” dimension.

From a List of files ( the most useful one…):

  • A list of netcdf files or grads files can be provided.
  • Time aggregation will be made by concatenating all time dimension in the list. Time does not need to be contiguous.

From Matlab data file:

  • Lat/Lon uniform grid will be created depending on the array dimension length.
  • Time aggregation is performed on time dimension length.
  • Array dimension order assumption is (t,y,x)

The CMOR Data Converter code is run from the command line on a system with the required supporting libraries and input data files.


For more information on the CMOR Data Converter, contact Denis.Nadeau