The ERA5 replica at NCI follows the structure and naming convention provided by ECMWF wherever possible and can be found on NCI's global filesystem ('gdata').
Within each ERA5 project code, you will find the data organised by the following structure:
- Sub-collection
- Dataset
- Product-type
- Parameter
- Year
For example, the full path of a file within the collection will take one of the following forms:
ERA5-preliminary
/g/data/<project-code>/<sub-collection>/<dataset>/<product-type>/<parameter>/<year>/<file.nc>
ERA5-derived
/g/data/<project-code>/<sub-collection>/<dataset>/<version>/<product-type>/<parameter>/<file.nc>
The following table provides examples of the possible options within these fields:
Category | Value |
---|---|
ERA5 project code | e.g., 'rt52', 'zz93', 'uc16' (see table) |
Sub-collection | era5 era5-1 era5-land era5-derived era5-preliminary |
Dataset | single-levels pressure-levels cems wfde5 |
Product Type | reanalysis* monthly-averaged** monthly-averaged-by-hour*** |
Parameter | NCI uses the parameter 'shortName'. Parameter lists can be found in the ECMWF ERA5 documentation: *Please see ERA5 Parameter Shortname Exceptions page for known exceptions. |
Year | <YYYY> format |
Filename | parameter_sub-collection_product-type_dataset_YYYYMMDD-YYYYMMDD.nc parameter_sub-collection_dataset_version_YYYYMMDD-YYYYMMDD.nc (era5-derived cesm data) parameter_sub-collection_dataset_version_product-type.nc (era5-derived wfde5 data) |
* High-resolution sub-daily reanalysis data, also referred to as 'oper' or 'wave' in the case of wave data.
** Monthly-averaged data, also referred to as 'moda' or 'wamo' in the case of wave data.
*** Monthly-averaged-by-hour, also referred to as 'mnth' or 'wamd' in the case of wave data.