Forge pulls forcing data from open datasets (documented in roms-tools). Most access is automatic and anonymous — ERA5 atmospheric forcing, for example, needs no account at all. Two sources require a one-time (free) registration, each with a small setup step afterwards.
Copernicus Marine (GLORYS)¶
GLORYS ocean-state data provides initial and boundary conditions, and Forge downloads it on demand through the Copernicus Marine toolbox. Required for essentially every domain, including the Getting Started example.
Authenticate once, using the
copernicusmarineCLI that ships with Forge’s dependencies:copernicusmarine login # prompts for your username and passwordThis stores a credentials file under
~/.copernicusmarine/that every future run picks up automatically.
For non-interactive contexts (HPC batch jobs, CI), set the environment
variables COPERNICUSMARINE_SERVICE_USERNAME and
COPERNICUSMARINE_SERVICE_PASSWORD instead of running the login step.
TPXO (tidal forcing)¶
TPXO tidal constituents are needed only for domains with tidal forcing — not for the Getting Started example. TPXO’s license doesn’t permit automated fetching, so Forge cannot download it for you.
You’ll receive download instructions from the TPXO team. Download the TPXO10.v2a netCDF files and place them under your source-data directory in exactly this layout:
<source_data>/TPXO/TPXO10.v2a/grid_tpxo10v2a.nc <source_data>/TPXO/TPXO10.v2a/h_tpxo10.v2a.nc <source_data>/TPXO/TPXO10.v2a/u_tpxo10.v2a.nc
Find your source_data directory with python -m cstar_forge.config show-paths. If anything is missing at processing time, Forge fails with a
message listing the exact expected paths.
Other user-staged datasets¶
A few optional datasets follow the same “you download, Forge verifies”
pattern as TPXO — for example WOA climatology, GLOFAS river discharge, and
EMOD topography. Each handler checks a documented location under
<source_data>/ and raises with instructions if files are missing; see the
SourceData documentation for the full inventory.