UTM datum ED50 conversion discrepancies

[chrisveness]

UTM datum ED50 conversion discrepancies

[chrisveness]

Hmm. There are a couple of issues here.

The link you gave, epsg.io/transform#s_srs=23030&t_srs=4326&x=572120&y=6382750, specifies 'EPSG:4326 WGS 84' as the output coordinate system.

If you specify 'EPSG:4230 ED50' as the output coordinate system – e.g. epsg.io/transform#s_srs=23030&t_srs=4230&x=572120&y=6382750 – the latitude/longitude values match the ones you were seeing in the code above.

To get WGS84 latitude/longitude from the geodesy library, you might do as follows:

import Utm, { LatLon } from './utm.js';
const utm = new Utm(30, 'N', 572120, 6382750, LatLon.datums.ED50);
const latlonED50 = utm.toLatLon();
const latlonWGS84 = latlonED50.convertDatum(LatLon.datums.WGS84);
console.log('ED50 ', latlonED50.toString('d', 7));  // 57.5801190°N, 001.7939918°W
console.log('WGS84', latlonWGS84.toString('d', 7)); // 57.5794383°N, 001.7956491°W

This still leaves a difference between the geodesy library values and the epsg.io values of close to 3 metres. As epsg.io is closed-source(?), I can't check the cause of this.

However, in order to do this conversion, I had to change utm.js to import LatLonEllipsoidal from latlon-ellipsoidal-datum.js, rather than latlon-ellipsoidal.js. I will change this in the library.

[qwilka]

OK thanks. So part of the problem is that I omitted to convert the LatLon from ED50 to WGS84 in my geodesy code. Since you are going to update geodesy to facilitate the LatLon datum conversion, and since the discrepancy in the result between geodesy and epsg.io has been noted, I would propose to close this issue.