North American Regional Reanalysis (NARR) for Arctic Research: Dataset Issues


  1. 07/11/2016: All data has been converted from the original GrIBS to netCDF4. See our netCDF4 webpage. Some values have changed due to changes at the source. There are also some known errors in the current gribs which NCEP is looking at. We tried to set these known errors to missing.
  2. ~07/10/2015, data from April 2009 had been replaced at PSL as NCEP has rerun the mode from April 1, 2009 to January 31, 2015 to fix codes and Mexico precipitation data preparation. Info http://rda.ucar.edu/datasets/ds608.0/docs/rr4.pdf. Please redownload files if they cover that time period.
  3. Nov 1 2013: 2m air temperature, 10m winds for 2007 have been replaced because they were found to be inadvertently based on the 30m level and not the 2m and 10m levels. Please replace any of those files you obtained before that date.
  4. Model was rerun at NCEP due to an assimilation problem starting at Jan 2003 (details) and we have updated our files.
    If you have acquired surface level data prior to Feb 12, 2006 please retrieve the latest version as a library error caused the files to have some erroneous values.

  5. Extreme values in the NCEP GRIBs 12-21Z 10 Feb 2004, 0-9Z 24 Feb 2004, and 12-21Z 10 Mar 2004 for variables
    3-hourly Water Condensate Flux Convergence Accumulation (wcconv) and
    3-hourly Water Vapor Flux Convergence Accumulation (wvconv).