S-RIP zonal mean data set of global atmospheric reanalyses on pressure levels

About this page
This page provides instructions on how to access the data set of zonal mean quantities (and latitude-longitude data on some selected levels) prepared for Chapters 5 and 6 of the S-RIP. At the moment, the data set is accessible on a server at McGill University.

Citing this data set
A reference paper will eventually be written about this data set circa middle or end of 2017.


  • A new version (v3) of the data set is in production. Some minor technical changes were made to the data format (e.g., 90S and 90N are now included). What appears to be a glitch in JRA-55 and JRA-55C in October 2010 will also be corrected.
  • July 2017 Adding ERA5 (2010-2016)
  • July 2017 Extending MERRA2 to December 2016
  • June 2017 NCEP-NCAR and NCEP-CFSR are being re-processed. Vertical velocity will be set to NAN instead of 0 above 100 hPa in NCEP-NCAR. The vertical dimension of NCEP-CFSR flips when the transition from v1 to v2 occurs (in 2010). It is going to be corrected in the new version.
  • January 2017 -> Will update Era-Interim to 2016
  • December 2016 A new horizontal data set is made available on both the original grid and the common grid (link). Missing a variable or pressure level of interest? Please let me know through the contact form.
  • December 2016 A new version of the data set is being developed (v2) with a new NetCDF structure. The old version will not be updated anymore. Please let me know if you have questions about the new data.
  • October 2016 -> Processing ERA-20c
  • August 2016 -> MERRA2 stops in 2011, I have to wait for the data to be available after that date.
  • July 2016 -> MERRA and ERA-40 data was reprocessed to fix missing data.
  • October 2015 -> I am told some levels are missing in MERRA, especially the pressure levels close to the surface. I think the issue comes from the fact that it may not interpolate data below topography (http://gmao.gsfc.nasa.gov/research/merra/pressure_surface.php) and therefore the zonal mean is not computed. I will investigate.
  • October 2015 -> I was notified of some missing data in ERA-40 before 1979, I will have to investigate [fixed]
  • October 2015 -> Many reanalyses are being extended to Dec 2012, J-25, MERRA will be extended in the future
  • Mar 24th 2015 -> Original grid data set was created.
  • Mar 1st 2014 -> Truncation error was corrected in NCEP’s reanalyses.
  • Jan 27th 2014 -> Residual circulation was added.
  • Jan 26th 2014 -> The description of some variables in the NetCdf files was corrected. The data itself was not affected.

Verified variables
To ensure the quality of the data set, I appreciate when the derived quantities I provide are confirmed independently. So far, the following variables were confirmed:

  • U (zonal-mean zonal wind) in the common_grid data set (Thanks to Amy Butler)
  • EP fluxes (Thanks to Andrew Orr)
  • Residual circulation (Thanks to Thomas Birner)

If you had the chance to compare the values provided in this data set with other values computed independently, please let me know. It will contribute a great deal to insure the quality of our shared data set.

The two data sets

  • A comparison data set (common_grid) using a common grid for all reanalysis data sets (link to the data)
    All diagnostics are computed and provided on a common grid. Only pressure levels that are common to all reanalysis data sets are kept. Horizontal fields are interpolated to a 2.5 by 2.5 grid before any other computations are made.
  • A research data set (original_grid) using higher resolution (link to the data)

Downloading using bash script
You may use this script to download the NetCdf files. If using a bash script to download the data set, please use “sleep 2” in between each file to reduce the load on the server. The bash script can be customized to download specific reanalysis, grid type, file type and years. More information on file types and grid type is available in the next sections.

declare -a reanalysis=("ERA5" "20th_Century_Reanalysis_v2" "20th_Century_Reanalysis_v2c" "era_20c" "era_40" "era_interim" "jra_25" "jra_55" "jra_55AMIP" "jra_55c" "nasa_merra" "nasa_merra2" "ncep_cfsr" "ncep_doe" "ncep_ncar")
#grid='original_grid_v2' #on the grid I have downloaded for each reanalysis
grid='common_grid_v2' #on a 2.5 by 2.5 grid
#type='MOM_4xdaily_' #momentum diagnostics
#type='TEM_4xdaily_' #transformed Eulerian mean diagnostics
#type='ZMD_monthly_' #zonal-mean variables, monthly-mean data (also available for other file types)
type='ZMD_4xdaily_' #zonal-mean variables, six-hourly data
for year in {1979..1980}
for month in $(seq -f "%02g" 1 12)
for dataset in "${reanalysis[@]}"
        echo $url
        wget -Nx $url
        sleep 2

Description of data set and variables
The variables provided as part of the S-RIP zonal-mean data set are described in the following document. The NetCdf files are fully annotated.

Download (PDF, 154KB)

Useful links
S-RIP official webpage
SPARC official webpage

Contacting the author
Please let me know if anything is wrong with the data set or if the documentation needs corrections or clarifications. If your comments/questions may benefit other users, I suggest you post a reply directly on this page.

Leave a Reply

Your email address will not be published. Required fields are marked *