Maps of the French Artillery Survey Group, 1917 - ca. 1919

[Michael Gurley]

As part of my research into my great-grandfather, who served in the MG Co, 104th Infantry Regiment, 26th (Yankee) Division, I'm studying this particular trench map that covers the Foret D'Apremont.

 

The full resolution copy can be viewed/downloaded at:

https://catalog.archives.gov/id/6997174

 

My intention is to georeference the image using ArcGIS software and digitize the full map features as vector data layers.  I've been visually creating link points between the trench map and modern aerial photography base layer with reasonable/limited success.

 

My next attempt was to utilize the gridlines and tic marks that exist on the source document, but I'm having a difficult time determining what exact projection is used.  The margin clearly specifies Lambert, but when I designate my software data frame in various Lambert Coordinate Systems, the cursor/position values do not match up to the grid values shown.

 

The grid appears to be 100x100m cells.

 

Perhaps this is going to require a custom geographic projection.  My software might only include "modern" projections.

 

Has anyone out there undertaken a similar effort with ESRI GIS software and tools?

 

Thanks in-advance for any advice.

 

~ Michael

[JOVE23]

I can't provide any help for you but I hope you can crack this nut, I have a burial on a Lambert map that I'd love to sort out in Lat/Long.

[Howard]

The differences between converters are owing to

1. The  figure of the Earth used

2. The Datum

3. Exact projection used- i.e. which specific kind of Lambert projection

4. Meridian used, London St. Paul's, London Greenwich, Brussels Observatory or Paris

The trouble is, it is not clear which ones they used. Col. Jack's Report on Survey on the Western Front has some detail of the French Lambert maps but some remains elusive. I tried to crack it a few years ago and gave up so I would be very grateful if more knowledgeable people could fill in the details. In Col. Jack's report, mention is made of numerous errors in their tabulation of geodetic results, it may of course be entirely possible that the maps are in fact permanently in error and therefore not possible to geo-reference by "proper" means but of course, that is only speculation.

It is implied that the French used the Nord de Guerre datum which references the Plessis 1817 ellipsoid but none of the converters I have tried can be set to those. A look at the excellent French Plans Directeur maps may also help, one may have projection & datum information not included on the others. Late in the war, the British started to use the French Lambert projection and maps started to appear in that projection but the war finished before the changeover was properly underway- this means of course that a document must exist somewhere that contains the (accurate) projection information. TNA? I have not looked.

If you did find out by any means which parameters they used, you then have to find the 3 term or 7 term (e.g. Molodensky) conversion factors or calculate them yourself. When you have done that, you can write or use a converter to WGS84. I would be interested to see any results on this as all my efforts failed.

The best method so far is to fit the map to a modern map by using 3 or better, 4 known points. That way you can achieve a reasonable fit. If you compare the fit you can get using that method with the really excellent geo-referencing you can find on the National Library of Scotland maps page, you can see that the approach works. Most geo-referenced old maps "fit where they touch". Sometimes.

Howard

 

 

 

[WhiteStarLine]

Michael has been offline for a month, but has raised some difficult questions.

 

The Belgian NGI have published some conversion tools at http://www.ngi.be/FR/FR4-4.shtm.  A version from 5 years ago had some Carte de France conversions and the current one includes Lambert.  Not sure how relevant they are, but it also has some delta x and delta y values.

[JOVE23]

Ok, I lied, I can help a bit.

 

This info comes from Bill Rankin, who wrote a book about mapping called "After the Map: Cartography, Navigation, and the Transformation of Territory in the Twentieth Century." I e-mailed him last night and had a discussion about these crazy Lambert maps and he provided me with the following info:

 

Quote

The système Lambert projection is a Lambert Conformal Conic with standard parallels at 47.7° N and 52.3° N, central meridian at 7.737229° E, latitude of origin 49.5° N, false easting 500000, and false northing 300000. I got these values from the French survey reports published just after the war (and converting French grades to degrees).