How to convert coordinates in MGRS (Indian Thailand) to UTM (WGS 84) in ArcGIS 9.3.1?

[ChristinRichter]

Dear All,

I hope somebody can help me with this transformation problem.

I got a data set of GPS coordinates which were recorded in N-Thailand, and the problem is that they are in MGRS (datum: Indian Thailand, which is only used by Garmin and I think in GIS programs it is similar to Indian 1975), but I need to transform them in UTM coordinates with the datum WGS 84.

The UTM zone needed is 47 Q, and the MGRS zone is 47 QQU.

The data I have look for example like this:
X_MGRS       Y_MGRS
78854  20424
79337  20831
79269  20793

I tried many different ways to upload them in ArcGIS, but all failed. I must admit I never worked with MGRS data before, only with UTM or Lan/Lon data???..

First, I tried in ArcCatalog to ???Create feature class from XY Table??? (from the Excel file, using MGRS coordinates as above) and set the input coordinate system to ???Indian 1975??? and the output coordinate system to ???WGS 1984 UTM Zone 47N???, so it should transform it. It created a shape file, but didn???t display any points. As I thought it might be a problem with the MGRS coordinates, I recalculated them as UTM (by adding a ???7??? infront of the 5 digit code for the X coordinates, and an ???18??? infront of the Y coordinates). So I tried it again with ???Create feature class from XY Table???, this time with the UTM coordinates, but again the same problem that it can???t display them.

Second, still in ArcCatalog with ???Create feature class from XY Table???, I tried to use the MGRS coordinates and this time set the output coordinate system to ???Indian 1975???, because I thought maybe it has problems with the transformation between the datums. When I added the created shape file to ArcMap, a new window opened and asked for transformation, so I clicked on convert from Indian 1975 to WGS 84. However it was still not showing any points because in only changed the coordinate system in the ???data frame properties???, rather than for the shape file.

Third, I tried with Military Analysist. So in Military Analyst Toolbox, I chose ???Conversion???: ???Convert Coordinates in File???. For input coordinate system I selected MGRS (written in 1 cell, in the format 47QQU7885420424), for input datum I selected ???Indian 1975???, for output datum ???WGS 84???, for the output specifications I activated ???add output fields??? and ???UTM fields??? and create new output file. So it created a .dbf file which I added to ArcMap, I did ???display XY data???, selected the new coordinates, specified them as being in WGS 84 and then it could display the points, but the new column added by Spatial Analysis with the UTM coordinates showed ???47 R 778520 2820730???, but the truth would be???47 Q 778512 1820736???, so it basically thinks we are in the wrong subzone (R instead of Q) and the Y coordinate is also very different??????  Anyway, when I tried to export it as a shape file an error message appeared and then when I checked again the attribute table, it was suddenly completely empty, so now of course it couldn???t display any points anymore.

Fourth, instead of using an Excel file as the data source, I saved the Excel file as .dbf and did the same steps as above (first and second trial), but this also didn't show any success.

By now I???m completely lost and have no idea anymore how I could solve the problem.

I really would appreciate any suggestion or response.

Thanks a lot,
Christin

[MelitaKennedy]

Hi Christin,

I think this might be the easiest to modify:

....
First, I tried in ArcCatalog to �??Create feature class from XY Table�?� (from the Excel file, using MGRS coordinates as above) and set the input coordinate system to �??Indian 1975�?� and the output coordinate system to �??WGS 1984 UTM Zone 47N�?�, so it should transform it. It created a shape file, but didn�??t display any points. As I thought it might be a problem with the MGRS coordinates, I recalculated them as UTM (by adding a �??7�?� infront of the 5 digit code for the X coordinates, and an �??18�?� infront of the Y coordinates). So I tried it again with �??Create feature class from XY Table�?�, this time with the UTM coordinates, but again the same problem that it can�??t display them.
....

Add the digits as before to create UTM coordinates, but when you use the Create Feature Class from XY Table, or use Add XY Data, set the input coordinate system as Indian 1975 UTM Zone 47N. At that point confirm they're in the correct locations, then use the Project Tool to create a new dataset in WGS 1984 lat/lon or WGS 1984 UTM Zone 47N. For the geographic/datum transformation from Indian 1975 to WGS 1984, try Indian_1975_To_WGS_1984_4 if it's available, or Indian_1975_To_WGS_1984_2 if it's not.

Melita

[ChristinRichter]

Dear Melita,

thank you very much for your reply and help. Finally it worked! The first problem was that I didn't know that "Indian 1975 UTM Zone 47 N" existed as input coordinate system because I couldn't find it in the folders (apparently I didn't check in the "Other GCS" folder), and second I assumed a direct transformation by specifying a different output coordinate system than the input coordinate system would work, but it doesn't. But it worked at least with the "Project" Tool in ArcToolbox.

However, I have a new question now. Using the Project Tool, there are four possible geographic transformations from Indian_1975_to_WGS_1984. I used number 4 as you suggested, and then I tried number 2. The outcome will be very different, and there is a difference between those two transformations of about 190m. I also tried number 1 and that produces a similar result like number 2 (difference of maybe 4m). So which transformation is the best, or in other words, which one represents the points more closely to the truth? In the internet I found that no. 1, 2 and 4 are valid for whole Thailand, and no. 3 specifically for Bangkok area. My data were collected in Phu Khieo Wildlife Sanctuary in N-Thailand. Do you think transformation no. 4 or no. 1/2 are better for this specific geographic area?

Thanks again and best wishes,
Christin

[MelitaKennedy]

Hi Christin,

#1 and #2 were defined by DMA/NIMA/NGA (US military group). They both have very similar parameter values and will give similar results, but #2 is an update of #1. I can't find as much information on #4 but it listed in EPSG has having a little better accuracy (3m as opposed to 5m) than the NGA-defined transformations. So I have to recommend #4.

Melita