North American Regional Climate Change Assessment Program
North American Regional Climate Change Assessment Program
    Home Users Handout
About Data
Contact Us
For PIs
For Users
    Access Data
  User Directory
Output Data Catalog
General Results
    NCEP-Driven RCM Runs
Climate Change Results
NARCCAP User Meeting Handout

NARCCAP Data Overview

NARCCAP data is stored in CF-compliant NetCDF format. Filenames are of the form

  • VariableName is the IPCC/CF-convention Variable Name as listed in the data tables
  • ModelName is the 4-character RCM identifier
  • Time is the starting time of the file, in the form YearMonthDayUTC.
Each file contains data for a single variable. For 3-D data, a file covers one month. For 2-D data, a file covers 5 years, ending in a year that ends in 0 or 5. (Files covering the beginning and ending of the run may not contain a full five years of data.)

So, for example, contains data for variable tas from the RegCM3 regional model starting at 03 UTC on January 1st, 1981 and ending at 00 UTC on January 1st, 1985.

Useful Links


NetCDF is a data format for storing array-oriented scientific data, particularly geophysical data. It is highly portable, widely-understood, and has built-in metadata that makes the files self-describing.

The main data array has associated coodinate arrays, xc, yc, and time, for each of its dimensions. Time is an unlimited or record dimension (important for some tools). Each variable in a NetCDF file can have attributes associated with it.

NetCDF Header -- produced by ncdump -h


Outputs contents of a netCDF file as ascii. From Unidata, part of the NetCDF libraries.

Useful options:
ncdump -h -- prints headers
ncdump -v variable -- prints a variable
ncdump -f n1,n2 -- specify precision of output
ncdump -l length -- specify line length of output

Example output: pr-ncdump.csv -- produced using ncdump -p 5,5 -l 2000


NCO is a package of command-line utilities for manipulating netCDF files. Each NCO command does one very specific kind of manipulation, but you can combine them to powerful effect.

There are a number of options common to most or all NCO commands. These include -v for specifying the variable and -d for choosing a dimensional slice. Some NCO commands make use of a "record" dimension. This needs to be the first dimension and is unlimited in extent. In NARCCAP data, this is always the time dimension.

NCO Documentation:
  • ncap and ncap2: netCDF Arithmetic Processors (examples) -- algebraic manipulation of data
  • ncatted: netCDF Attribute Editor (examples) -- modify metadata
  • ncbo: netCDF Binary Operator (includes ncadd, ncsubtract, ncmultiply, ncdivide) -- (examples) -- math involving two files
  • ncea: netCDF Ensemble Averager (examples) -- average across multiple input files
  • ncecat: netCDF Ensemble Concatenator (examples) -- combine files into a single record
  • ncflint: netCDF File Interpolator (examples) -- combine inputs via weighted interpolation
  • ncks: netCDF Kitchen Sink (examples) -- copies data to ascii or output file
  • ncpdq: netCDF Permute Dimensions Quickly, Pack Data Quietly (examples) -- rearrange dimensions or pack data
  • ncra: netCDF Record Averager (examples) -- average across time (record dimension)
  • ncrcat: netCDF Record Concatenator (examples) -- combine sequential files
  • ncrename: netCDF Renamer (examples) -- rename dimensions, variables, or attributes
  • ncwa: netCDF Weighted Averager (examples) -- weighted average over one file


seasonal-average -- Shellscript using NCO commands to generate seasonal climatology


NCL is an interpreted programming language for scientific analysis and visualization. It is designed with NetCDF in mind and allows easy and transparent access to NetCDF data files. It has a large and powerful plotting library oriented toward geophysical data.

NCL's syntax is FORTRAN-like. It supports the idea of coordinate variables and attributes associated with variables. Most parameters are passed to functions via "resource" attributes.


2d-time-plot.ncl Plot a timeslice from a file. Used in NARCCAP QC process. Example results.

Extensive reference and support materials are available on the NCL website:

Other Tools

There are a number of other tools for working with netCDF data.

CDATModular data analysis environment
CDOLike NCO, but for climate model data
FerretOceanographic visualization/analysis tool
IDLCommercial scientific analysis package
MatlabCommerical scientific analysis package
NumPyPython scientific computing libraries
PyNGLScientific visualization for Python
RStatistical computing language


Importing NARCCAP Data into GIS


Data Download Walkthrough