Coordinate Reference Systems

ANUGA works in a projected coordinate system — all domain coordinates are in metres, not degrees. A Geo_reference object records which coordinate system is in use and the local origin offset (xllcorner, yllcorner) that keeps the mesh coordinates numerically small.

Every SWW output file stores the complete geo-reference so downstream tools (QGIS, Python post-processing, etc.) can interpret the coordinates correctly.

—

Why a local origin?

Raw UTM or national-grid coordinates for a real site are typically large numbers (e.g. easting 308 500 m, northing 6 189 000 m for Cairns, Australia). Storing every mesh vertex relative to a local origin keeps the values small and avoids floating-point precision loss during computation.

Coordinates stored in a SWW file are relative to the geo-reference origin. To recover absolute coordinates, add xllcorner / yllcorner back to the mesh x / y arrays.

—

WGS84 UTM (the common case)

Most real-world ANUGA simulations use a WGS84 UTM zone. Pass zone and hemisphere (and optionally the local origin) to Geo_reference:

import anuga

# Cairns, Australia — UTM zone 55, southern hemisphere
geo_ref = anuga.Geo_reference(
    zone       = 55,
    xllcorner  = 363000.0,   # easting of domain origin (m)
    yllcorner  = 8132000.0,  # northing of domain origin (m)
    hemisphere = 'southern',
)

domain = anuga.create_domain_from_regions(
    bounding_polygon,
    boundary_tags,
    geo_reference = geo_ref,
)

The EPSG code is computed automatically from the zone and hemisphere and written to the SWW file:

Northern hemisphere: EPSG = 32600 + zone (e.g. zone 31N → EPSG 32631)
Southern hemisphere: EPSG = 32700 + zone (e.g. zone 55S → EPSG 32755)

>>> geo_ref.epsg
32755

Alternatively, supply the EPSG code directly and let ANUGA infer the zone and hemisphere:

geo_ref = anuga.Geo_reference(
    epsg      = 32755,
    xllcorner = 363000.0,
    yllcorner = 8132000.0,
)
>>> geo_ref.zone
55
>>> geo_ref.hemisphere
'southern'

—

National and non-UTM coordinate systems

Some countries use national projected CRS that do not follow the UTM zone structure. Common examples:

Country / region	EPSG code	CRS name
Netherlands	28992	Amersfoort / RD New
Great Britain	27700	OSGB36 / British National Grid
Germany (ETRS89 / UTM zone 32)	25832	ETRS89 / UTM zone 32N
GDA2020 MGA zone 56 (Australia)	7856	GDA2020 / MGA zone 56
WGS84 geographic (lat/lon)	4326	WGS84

These work fine with ANUGA — pass the EPSG code directly. If pyproj is installed, the datum, projection name, and false easting/northing are populated automatically from the EPSG definition:

# Dutch simulation in Rijksdriehoekstelsel (RD New)
geo_ref = anuga.Geo_reference(
    epsg      = 28992,
    xllcorner = 120000.0,   # easting relative to RD New origin (m)
    yllcorner = 480000.0,   # northing relative to RD New origin (m)
)

>>> geo_ref.epsg
28992
>>> geo_ref.projection
'Amersfoort / RD New'
>>> geo_ref.datum
'Amersfoort'
>>> geo_ref.false_easting, geo_ref.false_northing
(155000, 463000)
>>> repr(geo_ref)
'(crs=Amersfoort / RD New, easting=120000.000000, northing=480000.000000, epsg=28992)'
>>> geo_ref.is_located()
True

Note

ANUGA does not perform any reprojection — it treats all coordinates as a flat Cartesian grid in metres. Convert your input data to the target projected CRS before building the domain. Tools such as pyproj, GDAL, or QGIS can reproject shapefiles, rasters, and point clouds as needed.

—

Wavetank and hypothetical simulations

When no real geographic location is involved (e.g. a laboratory wave-tank or an idealised test case) simply omit zone and EPSG:

geo_ref = anuga.Geo_reference()   # zone=-1, no EPSG
geo_ref.is_located()              # False

The domain coordinates are then treated as an arbitrary Cartesian system with no connection to any geographic CRS.

—

Checking what is stored in a SWW file

The geo-reference metadata can be read directly from a SWW file:

from anuga.file.netcdf import NetCDFFile
from anuga.coordinate_transforms.geo_reference import Geo_reference
from anuga.config import netcdf_mode_r

fid = NetCDFFile('my_simulation.sww', netcdf_mode_r)
geo_ref = Geo_reference(NetCDFObject=fid)
fid.close()

print(geo_ref)          # e.g. (zone=55, ..., epsg=32755)
print(geo_ref.epsg)     # EPSG code stored in file
print(geo_ref.is_located())

—