Japan GNSS · 5 of 6 · checked against GSI
Coordinates and heights in Japan: what changed in 2025, and what has always been different
Three things trip up engineers arriving with good habits from elsewhere: the national height standard changed on 1 April 2025; the plane coordinate system has 19 zones and swaps the axes (X is north); and any old Japanese dataset may still be on a datum that is about 400 metres away. This page gives the facts and the EPSG codes, and points at the two tools on this site that apply them.
The datum: JGD2011, and the 2025 name change to JGD2024
Horizontal positions in Japan are expressed in the Japanese Geodetic Datum 2011 (JGD2011), defined on the GRS80 ellipsoid and tied to ITRF; it replaced JGD2000 after the 2011 Tohoku earthquake moved the country by metres. Geographic JGD2011 is EPSG:6668. Your receiver's WGS84 output agrees with it to well within a metre for most purposes, but for survey work transform explicitly; do not assume identity.
On 1 April 2025 GSI revised the nationwide elevation results and renamed the datum JGD2024. The important line, quoted from GSI's own Q&A, is that only elevation values were revised: latitude, longitude and plane rectangular coordinates did not change on that date. Elevations moved because the standard itself changed, from levelling-based heights to heights realised by satellite positioning plus a geoid model. Triangulation-point heights shifted by as much as +57 cm (Miyagi) and -67 cm (Hokkaido); levelling benchmarks by +23 to -40 cm depending on region. The epoch of the new heights is set to monitor crustal deformation going forward.
Practical consequence: a height you pull from a pre-2025 Japanese map or benchmark list and a height you derive today with GNSS plus Geoid 2024 can differ by tens of centimetres for reasons that have nothing to do with your receiver. Write down which standard every height in your project is on.
The geoid: Geoid 2024 Japan and its surroundings
The model is Geoid 2024 Japan and its surroundings (ジオイド2024日本とその周辺), a gravimetric geoid built from gravity data only, released in trial form in March 2024 and formally on 1 April 2025. It is distributed in the ISG 2.0 format and as a 250 m GML grid. Geoid heights over the main islands are typically 30 to 40 metres, so an ellipsoidal height straight from the receiver is that much too high. KUON MAP's coordinate panel and the free geoid endpoint on this site use this model.
The plane rectangular coordinate system: 19 zones, X north
For engineering and cadastral work Japan uses the Plane Rectangular Coordinate System (平面直角座標系), a set of 19 Transverse Mercator zones numbered I to XIX, each with its own origin and a scale factor of 0.9999 on the central meridian. Hokkaido alone spans three zones (XI, XII, XIII); Tokyo is zone IX; Okinawa XV and XVI; the outlying islands take the rest. The zone is chosen by prefecture and is written on every survey drawing.
The trap is the axes: X is northing and Y is easting, the reverse of the mathematical convention most GIS software assumes. EPSG codes for JGD2011 are 6669 (zone I) through 6687 (zone XIX), in order; when you import a Japanese CSV into a GIS, check the axis order before you check anything else. The general explanation, with UTM next to it, is in coordinates and height.
| Zone | Main area | EPSG (JGD2011) |
|---|---|---|
| IX | Tokyo, Kanagawa, Saitama, Chiba (mainland), Gunma, Tochigi, Ibaraki, Fukushima, Yamanashi, Shizuoka | 6677 |
| VI | Kyoto, Osaka, Nara, Mie, Fukui, Shiga, Wakayama | 6674 |
| XII | Hokkaido, central part including Sapporo and Obihiro | 6680 |
| XV | Okinawa main island | 6683 |
(Selected rows; the full zone table is on GSI's site and in any transformation library such as PROJ.)
The 400-metre trap: Tokyo Datum
Before 2002 Japan used the Tokyo Datum on the Bessel ellipsoid. Its latitudes and longitudes differ from JGD2000/2011 by roughly 400 metres on the ground. Old CAD drawings, old cadastral data and some municipal files are still on it, sometimes without saying so. If an imported dataset lands one block away from where the satellites put you, this is the first thing to suspect.
Sources
GSI, Q&A on the nationwide elevation revision (1 April 2025; JGD2024 name; only elevations revised; size of the shifts). GSI geoid page and the Geoid 2024 release (trial March 2024, formal April 2025). EPSG registry for JGD2011 codes. This site's field record of a 122 m base-height error in Tokachi is the worked example of why heights deserve their own line in a project log.