The short answer is WGS84, and the short answer is not enough to do anything with. A datum is not a single fixed thing — it is an ellipsoid plus a realisation, and WGS84 has been realised several times as the satellite constellation and the global tracking network improved. Alongside it sit NAD83, which is what most US mapping is actually on, and ITRF, which is the scientific frame and moves. This guide explains what that means in practice, and when the difference is small enough to ignore.
A datum is a decision about two things. First, an ellipsoid — a smooth mathematical surface chosen to approximate the shape of the Earth, defined by its semi-major axis and flattening. Second, an orientation and origin: where that ellipsoid sits and how it is turned relative to the real Earth. Latitude, longitude and ellipsoidal height are all measured against that surface, so a coordinate without a datum is like a height without a floor level.
Historically the second part mattered enormously. NAD27, for instance, was fitted to the North American continent using a single origin point at Meades Ranch in Kansas, so it fitted the United States beautifully and nowhere else. NAD83 moved to a geocentric origin, which is why coordinates shifted by up to 100 m or so when the datum changed — not because anyone had moved, but because the surface the coordinates were measured against had.
WGS84 is the datum the GPS constellation broadcasts. It was defined in 1984 and has been refined several times since — WGS84(G730), (G873), (G1150), (G1674), (G1762) and onward — each a new realisation that aligns WGS84 more precisely with the international terrestrial frames as tracking data improved. The ellipsoid has stayed the same; the alignment has tightened. Between the earliest realisations and current ones the difference is of the order of a metre, and between consecutive modern realisations it is centimetres.
For most work this granularity does not matter, because a hand-held receiver is not accurate to a metre anyway. It matters for network RTK and for anything measured to two centimetres, where the realisation and the epoch are both part of the coordinate's identity.
| Datum | Ellipsoid | Where it is used | Typical offset from WGS84 |
|---|---|---|---|
| WGS84 | WGS84 (a = 6,378,137 m, 1/f = 298.257223563) | GPS broadcast, worldwide navigation | Reference |
| NAD83 | GRS80 (a = 6,378,137 m, 1/f = 298.257222101) | US and Canadian mapping, State Plane, most national datasets | About 1–2 m, effectively fixed |
| ITRF | GRS80-compatible | Scientific positioning, global reference frames | Centimetres, changing over time |
| NAD27 | Clarke 1866 | Legacy US records before the 1980s | Up to about 100 m |
Two things are worth pulling out of that table. The first is that WGS84 and GRS80 have the same semi-major axis and differ only in the flattening, in the eleventh significant figure — so the ellipsoids are effectively identical for any terrestrial purpose. The second is that NAD83 and WGS84 are not the same frame even though their ellipsoids nearly coincide. They are aligned to different realisations and have drifted apart by roughly a metre, which is why a coordinate labelled NAD83 and the same coordinate labelled WGS84 describe points about a metre apart.
The Earth's surface is not stationary. Tectonic plates move at rates of centimetres per year — the Pacific plate off North America moves several centimetres a year, and even the stable interior of the continent drifts. A datum that is fixed to the Earth's crust therefore accumulates a difference against a datum fixed to the Earth's centre of mass, and that difference grows with time.
This is why modern coordinates sometimes carry an epoch alongside the datum, written as something like NAD83(2011) or ITRF2014, epoch 2010.000. A coordinate without an epoch is a position at an unstated moment, and over a decade the ambiguity can be several centimetres. For boundary work that is well inside tolerance; for deformation monitoring it is the signal.
The metre that does not matter, and the one that does. A metre of datum difference is invisible on a trail map and unacceptable in a boundary survey, a utility record or an as-built. The test is not how accurate the measurement was, it is whether two datasets are going to be overlaid or compared with each other. Mixing NAD83 and WGS84 in one overlay produces features that are offset in a systematic and thoroughly confusing way.
CoordCalc publishes free surveying and coordinate calculators. Results are for reference — confirm against the governing standard before construction or procurement. About · Privacy & cookies · Terms of use · Developers