Availability of data and file format

The data is divided into a series of volumes with a consistent

directory structure. Volumes are filled in the order data is submitted and

therefore a single volume may contain simulations from multiple

Climateprediction.net experiments.

The first volume in the dataset (vol-000) is a quality controlled subset of early Climateprediction.net data that is documented seperately here. This volume contains aproximately 2500 runs with a total size of 13Gb. Subsequent volumes will be an order of magnitude larger.

Each simulation is stored within a volume at a URL matching the following pattern:

ftp://ftp.ceda.ac.uk/badc/cpdn/volumes/vol-xxx/model/yy/result-name

Where xxx is a zero-padded volume number, model

identifies the type of Climateprediction.net simulation

(e.g. standard for the HadSM3 slab model experiment or

sulphur for the HadSM3 + sulphur cycle experiment) and

yy is a directory branch designed to limit the size of

individual directories. The directory branch yy is currently calculated

as the first 2 characters of the simulation's UM id, which is usually

also part of the result-name. However, this feature should not

be assumed as future volumes may use a different convention.

Within each simulation directory data is stored as a series of NetCDF files conforming to the CF metadata standard.