You are here: Home / Services / Data Management / Projects and Cooperations / Past Projects / Ensembles / technical details

technical details

how to construct the datafiles to be stored

information how to create metadata (e-mail of 2006-06-02)

information how to create data (e-mail of 2006-03-31)

Here comes the essential - technical - information for the data storage:

which data?

We will store variables of all ENSEMBLES-experiments (existing at this time: SRESA1B, SRESA2, SRESB1, 1PCTTO2X, 1PCTTO4X, 20C3M) of all existing runs. Monthly mean data of all variables are mandatory. Please take from the list of variables the information which time intervals for which variable is to be stored additionally.

the data has to be in netCDF-Format

You already know the procedure of converting your data from the files you submitted to PCMDI for IPCC-AR4. It will be very helpful for ENSEMBLES when you use the same convention (CMOR) as you used for those data. If you don't use CMOR, please have in mind to provide the attributes (global attributes, coordinate attributes and variable attributes) specified for CMOR (cmor_users_guide, pages 4,5) with your datasets.

data per file

Per file only one variable on one level.

Separate files for monthly mean, daily mean, diurnal cycle values and other frequencies where required.


The filenames have to be composed as follows:

  • centre/model acronym (e.g. MPEH5: Max-Planck-Institute/Echam5)1)

  • scenario acronym (e.g. 1PTO2X: 1PCTTO2X)1)

  • run number (e.g. 1: run 1)

  • time interval (MM:monthly mean, DM:daily mean, DC:diurnal cycle, 6H:6 hourly, 12h:12hourly)

  • variable acronym [with level value if not surface] (e.g. hur850: relative humidity, 850 hPa)2)

  • --> example: MPEH5_1PTO2X_1_MM_hur850

1)take it from the list of acronyms

2)The variable acronyms are the "short names" in the first column of the list of variables

daily means

Note in the comments of your datafiles from which values you calculated the daily averaged/accumulated values: did you use hourly/3-hourly values or every time step?

additional attribute

It could be helpful to use an additional optional attribute within the netCDF description of a variable which contains the original name you used for it in your model, for example:

float CLCT(time, lat, lon) ;

CLCT:units = "1" ;
CLCT:long_name = "Total cloud cover" ;
CLCT:standard_name = "cloud_area_fraction" ;

about the usage of the CF variable "land_cover"

"land_cover" is a new created CF variable. Here comes an example how to use this variable without the need of writing character variables into the field:


  byte land_cover(lat)


    land_cover:flag_values=0, 1, 2;

    land_cover:flag_meanings="needleleaf_tree shrub ice";



Document Actions