What follows is an overview of the PCMDI strategy to establish "reasonable" variable names for climate model and observational data.  Use of these naming conventions is strictly adhered to at PCMDI, and will be utilized for projects managed by PCMDI such as AMIP, CMIP and PMIP. The need for adherence to these conventions is essential for efficient managment of AMIP, a project for which PCMDI anticipates receiving over 1Tb of model output from over 30 modeling groups by the year 2000.

Variable names are created with naming constructs. The complete listing of PCMDI standard variable names is fully consistent with the naming constructs, and both are updated regularly. For convenience, the table of PCMDI standard variable names also includes standard units and metadata titles used by PCMDI.

This schema has been made public in response to a request made by several modeling groups. Those who are appreciative of the potential advantages of developing standards for the community may wish to adopt the PCMDI standards for their own use. Institutions choosing to collaborate with this effort should contact PCMDI amip@pcmdi.llnl.gov. Requests will be taken to continuously add new fields to this list as the demand arises.

Certain rules and guidelines have been established for the creation of PCMDI variable names. For example, 1) Keeping the variable names as short as possible is a high priority, 2) Only basic properties of the variable should be described, and thus any spatial or temporal averaging (e.g., zonal and monthly respectively) is identified by file names, not the variable names, and 3) The methodology must ensure that the listing of PCMDI standard variable names can easily be continuously updated. This third criteria is the most serious constraint given the diversity of categories which are expected to arise (e.g., from the various sub-models of the atmosphere, ocean, land surface, trace chemistry, etc.). At PCMDI it has thus been concluded that constructing variable names in the future will require flexibility to expand the listing in ways that cannot be anticipated. Thus, many of the guidelines developed are unique to each variable property (e.g., albedo, radiation, temperature, etc.). This enables much more versatility than attempting to create a comprehensive set of rigid rules that all variable names must adhere to. On the other hand, a concerted effort is made to ensure that there is as much consistency as possible throughout the various variable properties.  The variable property is identified by the variable name prefix. Current prefixes in use are listed in the following table: