Loading Data¶

bathy supports multiple data sources and formats for bathymetric data.

GEBCO OPeNDAP¶

You can download data directly from the GEBCO global bathymetry dataset via OPeNDAP:

import bathy

data = bathy.load_gebco_opendap(
    lon_range=(-12, -5),
    lat_range=(46, 50),
    year=2025,  # GEBCO version year
    save_path="data/region.nc",  # Optional: cache locally
)

Caching behaviour¶

If save_path is provided and the file already exists, bathy will load from the file instead of downloading. This avoids redundant downloads.

If save_path is omitted, data is downloaded to a temporary file that is automatically deleted after loading. For large regions, provide save_path to keep the file.

Download size¶

Download size scales with the requested area. The full GEBCO global grid is ~8 GB. bathy estimates the download size and logs it during download. For regions estimated at over 500 MB, save_path is required — this prevents accidentally filling the system temp directory with large files that are immediately deleted.

Valid years¶

GEBCO datasets are available for the following years: 2019-2025. An invalid year will raise a ValueError.

Preset regions¶

Common oceanographic regions are available as presets:

import bathy

# List all available regions
regions = bathy.list_regions()
print(regions[:10])
# ['arabian_sea', 'baltic_sea', 'bay_of_bengal', 'black_sea', ...]

# Use a preset
data = bathy.load_gebco_opendap(region="mediterranean")

EMODnet Bathymetry (European seas)¶

EMODnet provides high-resolution (~115 m) gridded bathymetry for European maritime areas via a Web Coverage Service (WCS):

import bathy

data = bathy.load_emodnet_wcs(
    lon_range=(-10, -5),
    lat_range=(50, 55),
    save_path="data/emodnet_region.tif",  # Optional: cache locally
)

This returns a GeoTIFF-backed xr.DataArray with lon/lat coordinates. The same preset regions and caching behaviour as GEBCO apply:

data = bathy.load_emodnet_wcs(region="north_sea")

Note

EMODnet coverage is limited to European seas. Requests outside this area will return an error.

The same download size estimation and save_path requirement for large regions applies to EMODnet downloads.

Local NetCDF files¶

Load from local NetCDF files (e.g., downloaded GEBCO data):

data = bathy.load_bathymetry(
    "path/to/gebco_2025.nc",
    lon_range=(-10, -5),
    lat_range=(50, 55),
    var_name="elevation",  # Variable name in file
    lon_name="lon",        # Longitude coordinate name
    lat_name="lat",        # Latitude coordinate name
)

GeoTIFF files¶

Load from GeoTIFF rasters:

data = bathy.load_bathymetry("path/to/bathymetry.tif")

GeoTIFF files are loaded using rioxarray. The CRS is read automatically:

Geographic CRS (e.g. EPSG:4326): coordinates are renamed to lon/lat
Projected CRS (e.g. UTM): coordinates are kept as x/y in metres

All analysis and plotting functions adapt to the CRS automatically. See Projected Coordinate Systems for details.

Clipping to a region¶

data is an xr.DataArray, so xarray's standard selection works directly:

subset = data.sel(lon=slice(-10, -7), lat=slice(47, 49))

Exporting data¶

To GeoTIFF¶

bathy.to_geotiff(data, "output.tif", crs="EPSG:4326")

To NetCDF¶

data.to_netcdf("output.nc")

Inspecting data¶

data is a standard xr.DataArray — use xarray directly:

print(data.shape)   # Grid dimensions
print(data.dims)    # ('lat', 'lon') or ('y', 'x')
print(data.coords)  # All coordinates
data                 # Rich repr in Jupyter