MGRS Datum transformation error

[DerekFrisby]

I am a Marine vet turned academic historian working with MGRS data from the Vietnam War. I have an overlay georeferenced 1:12500 map of Hue, Vietnam (NIMA Reference L909XHUE) with an imagery basemap. When I use ArcMAP MGRS geocoding, my feature points appear significantly off in the newly created feature layer.

When I use this mapplet to geocode, I get a spot on location. The mapplet appears to use the following datum to successfully plot the grid:

Projection: UTM

Datum: Indian 1975, Thailand

Ellipsoid: Everest (India 1830)

Equatorial Radius (m): 6377266.345

Polar Radius (m): 6356075.413

Eccentricity: .081472981

Flattening: 0.003324449

Datum Transformation (local to WGS84): dX (m) 210, dY (m) 814, dZ (m) 289

The mapplet produces the correct point from MGRS translation as follows:

DMS WGS84 LatLong is 16:28:18.87963, 107:35:39.54475

UTM 48Q 777400 1822600

For example, 48Q YD 764266 (Hue, Vietnam) should be at the edge of the Perfume River (known in 1968 as the LCU ramp). However, ArcMAP places it significantly SSE of the location just above a stadium. The incorrect ArcMAP geocoding data transformation (WGS84 or any similar Indian 1975, etc. datum I can find) displays as 48QYD7739522600  MGRS, 107°35'53.094"E  16°28'8.44" or 48Q 777395 1822599  UTM

I have tried without success to find/construct a custom datum in ArcMAP to replicate the above datum. Only mapplet datum plots correctly, but I need to geocode nearly 500 point features.

Any suggestions on how to build a custom datum or is their a native ArcGIS/ArcMAP datum I should use that would be identical?

Suggestions welcome. If you need any more info, please ask.

[ChrisDonohue__GISP]

Melita Kennedy (ESRI staff) might know the answer to this.

Chris Donohue, GISP

[MelitaKennedy]

Are you setting a transformation? The equivalent transformation In ArcMap to the one specified in the mapplet is Indian_1975_To_WGS_1984_2. We could also be having an issue between whether the "new" or "old" MGRS style is being used.

[DerekFrisby]

Melita,

I think you may be on to something. However, as a history professor learning GIS and an ArcMap novice, I could not find WGS_1984_2 or NAD27 datum in the ArcMap program. I am guessing that since there are many WGS_1984 versions that I am missing/overlooking one, and similarly that NAD27 is NAD 1927, but I have no idea which one to use.

I am also guessing a user would use the Projection tool--Convert Coordinate Notation. Then using my csv file, select Indian_1975 for the input coordinate system, but using the WGS_1984_2 as the output file?

OR

would use the Convert Coordinate Notation from Indian_1975 to NAD27, then NAD27 to WGS_1984_2

I know the NAD27 was for the "old" MGRS but I wouldn't know where to begin on translating it to the "new"

Based on the current feature point known correct location and the current projection incorrect location in my arcMap, I am guessing the NAD27 might be a possibility. Any idea on how to, if necessary, adjust for this in the transformation?

PS - Thanks Chris! Melita was an excellent contact.

[DerekFrisby]

I found this Web Geographic Translator app from the National Geospatial-Intelligence Agency (NGA) that seems to work.

It transforms INH-A1 - Indian, 1975 Thailand (Ellipse: Everest 1830) MGRS data to WGS_84 UTM data with very accurate results.

Is there a INH-A1 datum that ArcMap can utilize to geocode addresses?

[DerekFrisby]

​INH-A1 seems to be the same as Indian 1975 according to another site. apologies. Still confused why geocoding is off the mark. I know I am missing something simple.

[MelitaKennedy]

More than NAD27 used the 'old' (alternative) MGRS lettering schema. It is ellipsoid-based so Bessel 1841 was one of them, for instance. Although I would expect it to offset even more.

If you try the Convert Coordinate Notation Tool,

Input coordinate system would be GCS_Indian_1975 and output would GCS_WGS_1984.

Input coordinate format would be MGRS and output coordinate format probably DD_2.

Before you open the tool though, right-click ArcToolbox in the ArcToolbox window and open Environments.

Open the Output Coordinates section. In the Geographic Transformations pull-down, select Indian_1975_to_WGS_1984_2 (you can start typing the name and it will scroll the list). OK Environments window, then open the Convert Coord Notation Tool.

Melita

[DerekFrisby]

I really am perplexed by this. While helpful, I don't think the steps worked, or I performed them incorrectly.

I did notice that the georeferenced map I placed over the basemap has GCS_WGS_1984_Web_Mercator_Auxillary_Sphere

as its spatial reference in layer properties.

Still, 48QYD774226 should come up near feature #101 (LCU Ramp on the Perfume River bank) on the historical map not near the stadium (feature #149)

Here is the map package if you want to see what is what.

Any suggestions?

Hue_City_1968GR.mpk - Google Drive

[MelitaKennedy]

I think I see. The image is georeferenced to Web Mercator which is based upon a spherical WGS84. You're not going to be able to match the MGRS/UTM values on the map because they're using a different datum.

If I take 48QYD774226, unproject it to Indian 1975 latitude/longitude (16.47033139 107.59849798), transform it with 1304 (16.47191094 107.59431847), then convert it back to MGRS based on WGS84: 48QYD769229

That matches nicely with using the information button in the area of the boat ramp.

I hope this is helpful, and makes sense.

Melita

[DerekFrisby]

It is very helpful and sounds like the solution. I just wish I knew exactly how to replicate your solution.

I removed the georeferenced map from the table of contents. Then, do I go to the toolbox "Projections and Transformations" to unproject it or do I go to the Data Frame Properties--Coordinate System---Transformations?

I apologize for needing step-by-step instructions.