The Hybrid Coordinate Ocean Model (HYCOM) is an open-source data-assimilative computer model used to simulate and predict ocean circulation, temperature, salinity, and water velocity. These are Icechunks from the data on AWS Open Data.
# Dependencies for the HYCOM GOFS 3.1 virtual-Icechunk notebooks.
#
# pip install -r requirements.txt
#
# Lower bounds only, no lock file. The bounds are the versions the smoke-test
# notebook was last run against, in a clean Python 3.12 venv on 2026-09-18.
#
# REQUIRES PYTHON >= 3.12. Every icechunk 2.x release is published
# `requires_python = ">=3.12"`, so on an older interpreter
# `pip install "icechunk>=2.2"` resolves to nothing at all. Check `python -V`.
icechunk>=2.2
virtualizarr>=2.7 # ChunkManifest/ManifestArray and the Icechunk writer, incl. `region=` writes
zarr>=3.4
xarray>=2026.7
# Anonymous listing of the source bucket and the one-per-file header reads.
obstore>=0.11
numpy>=2.3
pandas>=2.2
# The documented read path is `xr.open_zarr(..., chunks={})`, which hands the arrays to
# dask; without it that call raises "chunk manager 'dask' is not available".
dask[array]>=2026.1
# No NetCDF parser is needed to BUILD the store: the source is NetCDF-3, and the
# references are computed from headers read by hycom_virtual.py itself. So there is no
# kerchunk, scipy, h5py or fsspec here. netCDF4 is used only to CHECK the store against
# an independent reader (its C library reads `<url>#mode=bytes` by byte range).
netCDF4>=1.7
boto3>=1.42 # Source Cooperative uploads: the viewer and the mirrored docs
compliance-checker>=6.1 # the CF check in the smoke test
matplotlib>=3.10 # the verification map