Origin of the War North
projection
Here is an excerpt of the thesis defended in
November 2005 by Nicolas Guilhot, currently Lecturer in History and Management
at IFROSS – University Lyon 3, at the Graduate School of Social and Law
Sciences of LYON II University – LUMIERE. Those
wishing to consult the full text will refer to: http://theses.univ-lyon2.fr/documents/lyon2/2005/guilhot_n
“As in many other areas of official
cartography, the first world war confirmed and accentuated the current
evolution in the projection systems used, and, even, if the preference for the
quality of conformity was not questioned, a projection system more suitable to
the new military needs was adopted: the war of position and the central role of
artillery fire according to maps had demonstrated the need for the military to
have, not only, geographical coordinates (degrees or grades), but also
rectangular (kilometric) in order to obtain the coordinates of a position
directly by their measurement on the map. However the polycentric system did
not allow assembling several sheets in a coherent system of rectangular
coordinates, a gap particularly problematic given the small area covered by the
breaks in the master plans. During the
war, the SGA had adopted Lambert’s projection in 1915, a consistent modified
conic projection.
In Lambert’s projection, the consistency i.e the
conservation of the angles, required a correction that increased with the
distance from the average parallel chosen for the projection, which made it
necessary to limit the area of use of the system in the north-south direction
to avoid excessive length alteration. The
Lambert system adopted during the war, called the war north, was chosen to make
an eastward extension for obvious military reasons: it could not, therefore, be
extended to the whole of France after the end of the conflict. However, since the use of rectangular coordinates proved
to be also advantageous from a topographic point of view, especially for the
determination of the planimetric pattern, the SGA conducted studies to adopt an
identical grid system on a large-scale map, 1: 20 000 or 1: 50 000, which led
to the adoption, in 1920, of new projection systems for the whole of France.
For the same military reasons, the Lambert
projection, indefinitely expandable to the east, was preferred to the Gaussian projection,
a particular case of the Lambert projection that can be indefinitely stretched in
the north-south direction. Due to the
increase in deformations with the distance from the reference parallel of the
projection, three different systems were adopted in France: Lambert north zone,
Lambert central zone, Lambert south zone, all referring to the ellipsoid of
Clarke (1880) and to the starting elements of the new triangulation
(geographical position of the Pantheon and Rosny azimuth). Tables and systems of transformation were later developed
to, gradually, harmonize the existing sheets, but the base maps could not be
retouched and only the coordinates indicated in the frame of the map and the
grid used were modified as long as the map design was not completely redone as
a result of new surveys. “
This
projection was used since the First World War until the end of the 1940s. It
covered part of northern France and, between the two wars; it was completed by
the adoption of Lambert I, II and III zones. The following figure, taken
from http://www.echodelta.net: The
principle of the “Modified British” coordinate system shows the zones corresponding to the projections in progress
during the Second World War.
The War North projection is a secant Lambert
projection (a single reference parallel), which uses the Du Plessis 1817
ellipsoid (instead of the Clarke 1880 ellipsoid used for Lambert I, II and
III), and which refers to to the ATF (Old French Triangulation) geodetic system
and not to the NFT (New French Triangulation) used by Lambert I, II and III
projections.
To integrate War North data into a GIS, it will take
a lot of patience and all the options available in your software for
customizing coordinate systems!
The management of
the War North projection with ArcGis
In the latest versions of ArcGis, you will find this
projection in the list of French screenings. However,
it has an error for False_Easting value which has to be 600,000 (not 500,000).
The problem is that there is no transformation
method available to switch from ATF to NFT or RGF93. In the case of the War North projection this results in a
shift of coordinates of several tens of kilometers. Therefore, it is necessary to create the two transformation
methods in order to integrate the War North data on Lambert layers (classical 93).
Once done, you should know that an inaccuracy of
about 100m will persist, with no hope of correction. 1/100 000 maps will have an offset of the order of one
millimeter, those with a scale 1/50000,
it will be of the order of 2 mm. Below 50,000 it
will be necessary to use other methods of referencing rather than the simple
transformation of coordinates.
The parameters of the War North projection are the
following:
False_Easting 600000.0
False_Northing 300000.0
Central_Meridian 6.00
Standard_Parallel_1 55.00
Scale_Factor 0.99950908000000
Latitude_Of_Origin 55.0
With regard to the geodetic system, the parameters
are:
Name: GCS_ATF_Paris
Angular Unit: Grad (0.01570796326794897)
Prime Meridian: Paris_RGS (2.337208333333333)
Datum: D_ATF
Spheroid: Plessis_1817
Semimajor Axis: 6376523.0
Semiminor Axis: 6355862.933255573
Inverse Flattening: 308.64
Setting up
transformations in ArcMap
The transformation of ATF to NTF or RGF93 is done
through a three-parameter transformation dX, dY, dZ.
The values to switch from ATF to NTF are:
dX = 1295,
dY = 82 and
dZ = -263
The values to change from ATF to RGF93 are:
dX = 1127,
dY = 22 and
dZ = 57
To set up the transformation of this geodetic system
to the one you usually use:
1-Open a new project in ArcMap
2-Load a layer in the usual projection system (Lambert
93 or Lambert I, II, II, IV, or II extended)
3-Load a map in War North projection
You will see the window telling you that the
geodetic systems are different and that offers you to set up a transformation.
4-Click the Transformations button If
you work in Lambert I, II, II, IV or II extended
5-Make sure that in Convert since you have GCS_ATF_Paris
6-That in To, you have GCS_NTF
7-Click on New
8- In Name type ATF_NTF
9 In parameters enter
the values 1295, -263 and 82 for, respectively, X, Y, Z.
10-Click OK , and OK
The next time you have to enter a transformation,
you will find this transformation in the drop-down list. If
you work in Lambert 93 or CC
5-Make sure that in Convert since you have GCS_ATF_Paris
6-That in To, you have GCS_RGF_1993
7-Click on New
8- In Name type ATF_RGF
9 In parameters
enter the 1127,22,57 values for, respectively, X, Y, Z.
10-Click OK , and OK
The next time you have to enter a transformation,
you will find this transformation in the drop-down list.
