Examining Simulation Results

The IRC model simulations (Century5, DayCent5, C5MPI, and IRC) will write results to files in either netCDF or spreadsheet format for a single site, or to a netCDF file with a spatial structure for multiple sites.

CSV Format

Spreadsheet output files are text files formatted as comma-separated values (CSV), which can be imported into most spreadsheet programs, or read by R, S-Plus, and other data analysis software.

The spreadsheet files have the extension "csv". Six monthly files are created by Century5, and seven by DayCent5. Also, DayCent5 produces three daily output files. The files are:

file name suffix	type	output set
*-co2.csv	monthly	CO2
*-cropC.csv	monthly	Crop and Grassland C
*-forestC.csv	monthly	Forest C
*-mtg.csv	monthly	(DayCent5 only) Trace gas fluxes
*-nps.csv	monthly	N, P, S
*-soilC.csv	monthly	Soil C
*-wt.csv	monthly	Soil Water and Temperature
*-dwt.csv	daily	(DayCent5 only) Soil Water and Temperature
*-gpp.csv	daily	(DayCent5 only) Maintenance Respiration
*-tg.csv	daily	(DayCent5 only) Trace gas fluxes

Available in Century5, DayCent5, C5MPI and IRC.

netCDF Format

The netCDF file format is a structured file format which is portable across computer platforms (transfer as "binary" file type), but cannot be viewed with a text viewer or editor. You will need a software application that can read netCDF files. In R you can use the ncfd package to read netCDF files.

The netCDF files have the extension "nc". Six monthly files are created by Century5, and seven by DayCent5. Also, DayCent5 produces three daily output files. The files are:
(To Do: table)

Available in Century5 or DayCent5, and C5MPI. Not available in IRC.

Spatial netCDF Format

The spatial netCDF format allows the results for multiple sites to be stored in one set of files. As with the other formats, there are six monthly files are created by Century5, and seven by DayCent5, and DayCent5 produces three daily output files; see the CSV format for the list of file name suffixes. Each site is identified by a unique site identification number. Also, each site can have its own latitude and longitude location, and site name. The names of the output variables are stored in the netCDF file.

The structure of the spatial netCDF files is shown below (using CO2 output as an example):

netcdf co2 {
dimensions:
  numVars = 45 ;
  numSites = 1 ;
  time = 73 ;
variables:
  float time(time) ;
    time:long_name = "time" ;
  float lat(numSites) ;
    lat:long_name = "latitude of center of cell" ;
  float lon(numSites) ;
    lon:long_name = "longitude of center of cell" ;
  uint id(numSites) ;
    id:long_name = "site parameter identifier" ;
  float values(time, numSites, numVars) ;
    values:long_name = "CO2 respiration" ;
    string values:VarNames = "AMT1C2", "AMT2C2", "AS11C2", "AS21C2", "AS2C2",
      "AS3C2", "AST1C2", "AST2C2", "CO2CCE(1,1,1)", "CO2CCE(1,1,2)", "CO2CCE(1,1,3)",
      "CO2CCE(1,2,1)", "CO2CCE(1,2,2)", "CO2CCE(1,2,3)", "CO2CCE(2,1,1)", "CO2CCE(2,1,2)",
      "CO2CCE(2,1,3)", "CO2CCE(2,2,1)", "CO2CCE(2,2,2)", "CO2CCE(2,2,3)", "CO2CPR(1)",
      "CO2CPR(2)", "CO2CRS(1)", "CO2CRS(2)", "CO2CTR(1)", "CO2CTR(2)", "MT1C2(1)",
      "MT1C2(2)", "MT2C2(1)", "MT2C2(2)", "RESP(1)", "RESP(2)", "RESPH", "S11C2(1)",
      "S11C2(2)", "S21C2(1)", "S21C2(2)", "S2C2(1)", "S2C2(2)", "S3C2(1)", "S3C2(2)",
      "ST1C2(1)", "ST1C2(2)", "ST2C2(1)", "ST2C2(2)" ;
}

This is an abbreviated view of the netCDF file header. To view all the header information, use the command "ncdump -h outputName-co2.nc". (Do not forget the "-h" or all the data will also be extracted!)

Not available in Century5 or DayCent5. Available in C5MPI. Planned availability in IRC.

For parallel (concurrent) netCDF performance, the HDF5 and netCDF libraries must be built with the parallel capability enabled. (See the document in the source code files doc/BuildEcosystemModels.html for details.)

Utilities and Scripts

In the models' source code directories at Ecosystem/Century5/scripts are R functions to read the CSV output files and to draw graphs of some output variables. R can also read the netCDF output using the ncdf4 package.