Usually, all layers of geographic data have the
description of the coordinate system. Unfortunately,
this is not always the case. Generally, the problem increases by not being able
to contact the producer of the data.
Problem diagnosis
A- You ask Arcmap to load the layer and you see a
warning message: “ The
following added data sources do not have spatial reference information.
This data can be displayed in ArcMap, but it cannot
be projected. “
If it is a vector data layer (points, lines, polygons),
the data must lack a defined coordinate system. In
the case of shapefiles, the PRJ file is missing.
B- You ask Arcmap to load the layer and, without receiving
any warning or error message, you do not see your new data, or it is in a
different place from where it should be.
If it’s a vector data layer (points, lines, polygons),
the data has a defined coordinate system, but it’s false. The PRJ file exists, but its contents are wrong.
Search procedure
In this article we try to supply you a guide to
determine the projection system of a GIS data layer, when it is unknown.
We have limited ourselves to the most commonly used systems in France by the
various organizations and administrations: Lambert and Lambert93 projections,
UTM projections and non-projected data (geographical latitude / longitude).
Of course, there are many others, used punctually.
But it is impossible to determine them without a truly
detective work.
The following maps show the different areas affected by
each type of projection system.
We assume that you know how to use ArcGis and that
you have other reference data that will allow you to compare and assess the results.
You can download the pdf document of this article on
the NASCA
website , by clicking here
.
You must find the
range of X and Y values in the data.
1- Start ArcMap with a new empty document
2- Add the data with the unknown coordinate system.
The data must not have a defined coordinate system.
In the case of shapefiles, it must not have PRJ files.
If there is one, rename it differently.
3- Right-click on the name of the layer in the table of
contents,
4- Click Properties to open the Layer Properties dialog
box.
5- Select the Source tab and examine the data range.
(In the upper part of the window)
The term coordinate system can be applied to data
expressed in decimal degrees (geographic coordinates) or a projected coordinate
system expressed in meters.
If the coordinates shown in the Range are expressed in
decimal degrees, they will range from -180 to +180 for longitudes ( Left
and Right values ) and between -90 and +90 for latitudes ( High
and Low values ). . The data system will
be searched in the section “ ArcGis Geographic
Coordinate Systems”. It remains to find the data geodetic system (Datum). (See further)
If the coordinates shown in the Range are of the
order of hundreds of thousands or millions, they are meters. The data system will be searched in the section “ ArcGis
Projected coordinate systems “, and it will also remain to find the data geodesic system (Datum).
How to find the
projection system
Here you will find a logical progression based on
the value range of your layer. In addition to
these values, we will use the data area against three maps. Map
N ° 1: NTF Lambert zones
Map N ° 2: Lambert 93 areas
Map N ° 3: UTM time zones
Y coordinates <1,000,000
—- X coordinates between 0 and 1 100 000
—— See map N ° 1
——- If your data is in Lambert I zone
——– Data is in NTF Lambert I zone, EPSG code
27561.
——- If your data is in Lambert II zone
——– Data is in NTF Lambert II area, EPSG code
27562. –
——- If your data is in Lambert III zone
——– Data is in NTF Lambert III area, EPSG code
27563.
—— If your data is in Lambert IV zone
——– Data is in NTF Lambert IV zone, EPSG code
27564.
Y coordinates between 1,000,000 and 2,000,000
—– X coordinates between 0 and 1 100 000
—— See map N ° 1
——- If your data is in Lambert I zone
——– Data is in NTF Lambert I carto, EPSG code
27571.
—– X coordinates between 1,000,000 and 2,500,000
—— See map N ° 2
——- If your data is in zone 1 (green)
——– Data is in Lambert CC42 RGF, code EPSG 3942.
Y coordinates between 2,000,000 and 3,000,000
—– X coordinates between 0 and 1 100 000
—— See map N ° 1
——- If your data is in Lambert II zone
——– The data is in NTF Lambert II carto, code EPSG
27572
——- If your data is outside the Lambert II zone
——– Data is in NTF Lambert II extended, EPSG code
27572.
—– X coordinates between 1,000,000 and 2,500,000
—— See map N ° 2
——- If your data is in zone 2 (blue)
——– Data is in Lambert CC43 RGF, code EPSG 3943.
Y coordinates between 3,000,000 and 4,000,000
—– X coordinates between 0 and 1 100 000
—— See map N ° 1
——- If your data is in Lambert III zone
——– Data is in NTF Lambert III carto, code EPSG
27573.
—– X coordinates between 1,000,000 and 2,500,000
—— See map N ° 2
——- If your data is in zone 3 (green)
——– Data is in Lambert CC44 RGF, code EPSG 3944.
Y coordinates between 4,100,000 and 4,300,000
—– X coordinates between 500,000 and 600,000
—— See map N ° 1
——- If your data is in Lambert IV zone
——– Data is in NTF Lambert IV carto, code EPSG
27574.
Y coordinates between 4,000,000 and 5,000,000
—– X coordinates between 200,000 and 750,000
—— See map N ° 3
——- If your data is in Time Zone 30
——– Data is in WGS84 UTM 30N, EPSG code 32630.
——- If your data is in Time Zone 31
——– Data is in WGS84 UTM 31N, EPSG code 32631.
——- If your data is in Time Zone 32
——– Data is in WGS84 UTM 32N, EPSG code 32632.
—– X coordinates between 1,000,000 and 2,500,000
—— See map N ° 2
——- If your data is in zone 4 (blue)
——– Data is in Lambert CC45 RGF, code EPSG 3945.
Y coordinates between 5,000,000 and 6,000,000
—– X coordinates between 200,000 and 750,000
—— See map N ° 3
——- If your data is in Time Zone 30
——– Data is in WGS84 UTM 30N, EPSG code 32630.
——- If your data is in Time Zone 31
——– Data is in WGS84 UTM 31N, EPSG code 32631.
——- If your data is in Time Zone 32
——– Data is in WGS84 UTM 32N, EPSG code 32632.
—– X coordinates between 1,000,000 and 2,500,000
—— See map N ° 2
——- If your data is in zone 5 (green)
——– Data is in Lambert CC46 RGF, EPSG code 3946.
Y coordinates between 6,000,000 and 7,000,000
—– X coordinates between 0 and 1 250 000
—— See map N ° 2
——- If your data IS in zone 6 (blue)
——- There are two possibilities, you will have to
test them against a
——- known layer to determine which is the right one.
——– Data may be in Lambert RGF 93, code EPSG
21546.
——–Or
——– Data may be in Lambert CC47 RGF, EPSG code
3947.
——- If your data is NOT in zone 6 (blue)
——– Data is in Lambert RGF 93, code EPSG 21546.
—– X coordinates between 1,250,000 and 2,500,000
—— See map N ° 2
——- If your data is in zone 6 (blue)
——– Data is in Lambert CC47 RGF, code EPSG 3947.
Y coordinates between 7,000,000 and 8,000,000
—– X coordinates between 1,000,000 and 2,500,000
—— See map N ° 2
——- If your data is in zone 7 (green)
——– Data is in Lambert CC48 RGF, code EPSG 3948.
Y coordinates between 8,000,000 and 9,000,000
—– X coordinates between 1,000,000 and 2,500,000
—— See map N ° 2
——- If your data is in zone 8 (blue)
——– Data is in Lambert CC49 RGF, code EPSG 3949.
Y coordinates between 9,000,000 and 10,000,000
—– X coordinates between 1,000,000 and 2,500,000
—— See map N ° 2
——- If your data is in zone 9 (green)
——– Data is in Lambert CC50 RGF, code EPSG 3950.
If you follow this procedure to find the coordinate
system for QGis, it’s time to go back to the article How
to find the coordinate system of a layer in QGis? by clicking here
How to find the
geodetic system (DATUM)
Once you have gone through the first two stages, you
have one last point to determine. Any location
system necessarily refers to a center of the Earth. Since the Earth is not a perfect sphere, and that it is
necessary to calculate its center, there are several ways to calculate it and, therefore,
several different ” centers
“. The
difference is not huge and until a few decades ago it was a rather theoretical
discussion, the resulting positioning difference being generally less than
300m.
Technically, a coordinate system is always
associated with a geodetic system.
The Lambert 1,2,3,4 and extended 2 projections are
always associated with the NTF system (Nouvelle triangulation française)
The Lambert 93 and CC42 to 50 projections are still
associated with the RGF system.
So, if in the previous step you have come to define one
of these projections, the work is finished.
For UTM projections 30 to 32, in theory, the
associated system is the WGS84 systems. But they
can also be associated with the Europe 50 system.
In the case of geographical (non-projected) data, it
is, also generally, associated with the WGS84 system, but can also be found
associated with the NTF or Europe 50 system.
How to find out?
It is necessary to have a data reference layer, with
the projection system defined correctly and above all, having a good accuracy
(detail).
In a new project in ArcMap, load this reference layer.
Open the properties of the layer-> Source and note
the geodetic system (DATUM) of this layer: you will find it at the very bottom
of the Data Source window.
In ArcCatalog, set the unknown layer coordinate system
as determined in the previous steps, using the most common geodetic system:
WGS84
Load the layer in ArcMap.
If you have no message,
-and the data appears in the
right place and there is no slight shift (100-300m), you’re done. The definition you have adopted is the right one.
-and you have an offset of all
your entities, of the order of 100 to 300m, your data is not in WGS84. The coordinate system (UTM or geographic) is good, but your
data is not in WGS84. It must be in Europe 50 or
NTF.
On
the other hand your reference data is in WGS84, if no message appears.
If your entities are shifted downwards (SW) by about
230m, your layer is in NTF.
If your entities are shifted upwards (NE) by about
130m, then your layer is in Europe 50.
Remove the ArcMap layer. In ArcCatalog go to Layer Properties and XY Coordinate
System, double click on the coordinate system. The
system edit window opens. Click edit and go to
” Geographic Coordinate Systems
»->« Europe
and click on either NTF (Paris) or Europe Datum
1950.
If you have a warning message,
Your
reference layer is not in WGS84. It is necessary to configure the transformations, if not,
even if everything seems correct, in fact it will be wrong.
Click on the transformations button, open the window «
using » and
select
NTF_Paris_To_WGS_1984 if you are in NTF
or
ED_1950_To_WGS_1984_1 if you are in Europe 50
If the data appears in the right place and there is
no slight shift (100-300m), you are done. The
definition you adopted (WGS84) is the right one.
If you are in NTF and the data is shifted by about
230m, the unknown layer is also in NTF.
If you are in NTF and the data is shifted about 80m,
the unknown layer is in Europe 50.
If you are in Europe 50 and the data is shifted by
about 130m, the unknown layer is also in Europe 50.
If you are in Europe 50 and the data is shifted by
about 80m, the unknown layer is in NTF.