Datum vs Coordinate System vs Projection

A datum or reference frame connects coordinates to the physical Earth. A coordinate system defines the axes, units and directions used to express positions. A map projection converts positions on a curved reference surface into a flat coordinate space. A coordinate reference system, or CRS, brings the necessary pieces together so coordinate values can be interpreted.

The terms overlap in everyday GIS conversation, but they are not interchangeable. Distinguishing them makes it easier to diagnose a shifted layer, choose a valid transformation and understand why two datasets with similar-looking coordinates may still be incompatible.

Term · Main question it answers · Example

Datum or reference frame — How is the coordinate model anchored to the Earth? — WGS 84, ETRS89, NAD83
Coordinate system — What axes and units express a position? — Longitude and latitude in degrees; easting and northing in metres
Projection — How are curved-surface positions converted to a plane? — Transverse Mercator, Lambert Conformal Conic, Equal Earth
CRS — What complete definition gives these coordinates meaning? — WGS 84 / UTM zone 32N

The datum provides the terrestrial reference

Historically, a geodetic datum described an ellipsoid and how it was positioned relative to the Earth. Modern terminology separates the abstract datum from a reference frame: a practical realisation based on the measured coordinates of stations, satellites or other control points.

This distinction matters because the Earth moves: tectonic plates shift, equipment and models improve, and global frames are updated. WGS 84 consequently has multiple realisations, so a high-accuracy coordinate is incomplete without knowing which realisation and, for a dynamic frame, the coordinate epoch.

National and continental frames may be fixed to the motion of a tectonic plate so coordinates remain stable relative to that region. At a small map scale, two modern frames may appear identical. In surveying, engineering or change detection, their difference can be material.

A datum is therefore more than the name of an ellipsoid. It establishes how the mathematical coordinate model relates to the Earth being measured.

The coordinate system defines the values

A coordinate system specifies axes, axis directions, units and the rules for assigning numbers to positions.

A geographic coordinate system commonly expresses longitude and latitude as angular values on an ellipsoid. A projected Cartesian coordinate system usually expresses easting and northing in metres or feet. A geocentric coordinate system uses three axes through the Earth's centre. A vertical coordinate system expresses height or depth relative to a defined surface.

Because the same datum can support several coordinate systems, WGS 84 positions may appear as longitude and latitude, Earth-centred x/y/z coordinates or projected eastings and northings. Saying that a dataset is “in WGS 84” may therefore identify only part of what a user needs to know.

Axis order is also part of the definition. Some formal geographic CRS definitions order latitude before longitude, while many GIS formats and APIs expect x/longitude first. A pair such as 4.35, 50.85 can describe Brussels or a very different place depending on the convention. The coordinates have not changed; their interpretation has.

The projection performs the curved-to-flat conversion

A map projection is a coordinate conversion governed by a method and parameters. It translates geographic coordinates on an ellipsoid or sphere into planar coordinates.

The method name is not a complete CRS. Transverse Mercator, for example, can be configured with different central meridians, scale factors, origins and false coordinates, which is why the UTM system contains many CRSs based on that one method. Lambert Conformal Conic can likewise be tuned using one or two standard parallels for different countries or regions.

Projection parameters determine where distortion is controlled. A central meridian placed through the study area can keep scale variation low nearby; the same method with a distant central meridian may perform poorly.

How the pieces form a projected CRS

Consider WGS 84 / UTM zone 32N. Its complete meaning includes:

the WGS 84 geodetic reference framework and ellipsoid;
an underlying geographic coordinate system;
the Transverse Mercator projection method;
zone-specific parameters, including a 9° east central meridian;
a Cartesian easting/northing coordinate system in metres;
false-coordinate conventions and an area of use.

An EPSG identifier packages those components in a form software can exchange reliably. It does not guarantee that the CRS is appropriate for the task; it guarantees that the definition can be identified.

This is why replacing a full CRS description with “UTM”, “Mercator” or “WGS 84” discards information. Each phrase names only part of the chain.

Three different problems that are often called a projection problem

The CRS label is missing or wrong

The coordinate values may be correct, but software does not know how to interpret them. Assigning the correct CRS restores that interpretation without changing the numbers. Assigning a guessed CRS can place the layer in a plausible but false location.

The datums or reference frames differ

Two layers may use the same projection method and units but remain offset because their terrestrial references differ. They require an appropriate coordinate transformation, which may depend on location, accuracy, grids and epoch.

The projection is unsuitable for the operation

Layers can align perfectly while a calculation is wrong for its purpose. Measuring area in Web Mercator is the classic example: the CRS is known and the computation succeeds, but changing scale makes the planar result unsuitable as ground area.

Naming the layer of failure leads to the right remedy. “The projection is wrong” is too vague to guide an auditable correction.

Assignment, conversion and transformation

These operations are related but distinct:

Assigning a CRS attaches the correct definition to existing coordinates. The coordinate values do not change.
A projection conversion changes coordinates between systems based on the same datum or reference frame, such as WGS 84 longitude/latitude to WGS 84 UTM.
A datum or reference-frame transformation changes coordinates between different terrestrial references.
Reprojection is a convenient GIS umbrella term that may include a conversion, a transformation or both.

When source CRS metadata is absent, no reliable reprojection can occur until the source coordinates have been identified. A target CRS does not supply the missing source meaning.

What to record in a reusable workflow

At minimum, preserve the source and target CRS identifiers or complete definitions, the coordinate operation selected, and any accuracy or epoch assumptions. If a grid file was required, record that dependency. If software chose an operation automatically, inspect which one it chose.

A spatial answer should not be separated from the reference information that makes it reproducible. Keeping the source, coordinate operation and resulting layer together allows another person to review more than the final visual alignment. They can see what the coordinates meant and how that meaning changed.

References

ISO 19111:2019 — Referencing by coordinates. Formal model for datums, reference frames, coordinate systems, CRSs and coordinate operations.
EPSG Guidance Notes. Authoritative guidance on CRS definitions and coordinate operations.
PROJ coordinate operations. Practical implementation documentation.