Backgroud:  AMIP simulations submitted to PCMDI have been "filtered"  via (LATS) in accordance to the AMIP data transmission guidelines.  This data is either in GRIB or netCDF, but organized as suited most effectively to the modeling group's data processing stream.   While it is not difficult to study a single simulation with the data in this form,  it is not in a form practical for distribution to the AMIP research community.   To alleviate this  problem, PCMDI shapes the data from each model into a common structure, thanks to the invent of a powerfull new capability developed by Charlie O'Connor.  This product adheres to a metadata convention that is rapidly gaining popularity within the climate research community, providing a host of useful related information for research.
 

Data specifications:

GDT compliant netCDF.   The GDT convention provides a host of useful associated variables  (e.g., weights_latitude and bounds_latitude) and documentation.    These data can be accessed by a variety of software tools, such as PCMDI's CDAT.

Monthly mean data is organized as one variable/year/file.  This means that the files containing 2D fields (e.g., OLR) are small (but not too small) whereas data from a high resolution model is large (but not too large).  These yearly files can be "spanned in time" with a variety of software tools.  The structure of high frequency (6h and daily) data is further refined to one variable/level/year/file.

Data availability:

AMIP data is available via ftp to researchers participating in approved AMIP diagnostic subprojects.   netCDF files are compressed (with the gzip utility) to facilitate data transmission.