PROJ.4 - Frequently Asked Questions

Where can I find the list of projections and their arguments?

There is no simple single location to find all the required information. The PostScript/PDF documents listed on the main PROJ.4 page under documentation are the authoritative source but projections and options are spread over several documents in a form more related to their order of implementation than anything else. The proj command itself can report the list of projections using the -lp, option, the list of ellipsoids with the -le option, the list of units with the -lu option, and the list of built-in datums with the -ld option.

The GeoTIFF Projections Pages include most of the common PROJ.4 projections, and a definition of the projection specific options for each.

How do I do datum shifts between NAD27 and NAD83?

While the nad2nad program can be used in some cases, the cs2cs is now the preferred mechanism. The following example demonstrates using the default shift parameters for NAD27 to NAD83:

% cs2cs +proj=latlong +datum=NAD27 +to +proj=latlong +datum=NAD83 
-117 30

producing:

117d0'2.901"W   30d0'0.407"N 0.000

In order for datum shifting to work properly the various grid shift files must be available. See below. More details are available in the General Parameters document.

How do I build/configure PROJ.4 to support datum shifting.

After downloading and unpacking the PROJ.4 source, also download and unpack the set of datum shift files. This would be a file like ftp://ftp.remotesensing.org/pub/proj/proj-datumgrid-1.3.zip. This file should be unpacked within the proj/nad directory. Then proceed with the configuration, build and install. This will ensure that the build system knows about the grid shift files, and applies the ascii to binary preprocessing step.

On Windows the extra nadshift target must be used. For instance nmake /f makefile.vc nadshift in the proj/src directory.

A default build and install on Unix will normally build knowledge of the directory where the grid shift files are installed into the PROJ.4 library (usually /usr/local/share/proj). On Windows the library is normally built thinking that C:\PROJ\NAD is the installed directory for the grid shift files. If the built in concept of the PROJ.4 data directory is incorrect, the PROJ_LIB environment can be defined with the correct directory.

How do I debug problems with NAD27/NAD83 datum shifting?

  1. Verify that you have the binary files (eg. /usr/local/share/proj/conus) installed on your system. If not, see the previous question.

  2. Try a datum shifting operation in relative isolation, such as with the cs2cs command listed above. Do you get reasonable results? If not it is likely the grid shift files aren't being found. Perhaps you need to define PROJ_LIB?

  3. The cs2cs command and the underlying pj_transform() API know how to do a grid shift as part of a more complex coordinate transformation; however, it is imperative that both the source and destination coordinate system be defined with appropriate datum information. That means that implicitly or explicitly there must be a +datum= clause, a +nadgrids= clause or a +towgs84= clause. For instance "cs2cs +proj=latlong +datum=NAD27 +to +proj=latlong +ellps=WGS84" won't work because defining the output coordinate system as using the ellipse WGS84 isn't the same as defining it to use the datum WGS84 (use +datum=WGS84). If either the input or output are not identified as having a datum, the datum shifting (and ellipsoid change) step is just quietly skipped!

  4. The PROJ_DEBUG environment can be defined (any value) to force extra output from the PROJ.4 library to stderr (the text console normally) with information on what data files are being opened and in some cases why a transformation fails. Note that PROJ_DEBUG support is not yet very mature in the PROJ.4 library.

  5. The "-v" flag to cs2cs can be useful in establishing more detail on what parameters being used internally for a coordinate system. This will include expanding the definition of +datum clause.

How do I use EPSG coordinate system codes with PROJ.4?

There is somewhat imperfect translation between 2d geographic and projected coordinate system codes and PROJ.4 descriptions of the coordinate system available in the epsg definition file that normally lives in the proj/nad directory. If installed (it is installed by default on Unix), it is possible to use EPSG numbers like this:
% cs2cs -v +init=epsg:26711
# ---- From Coordinate System ----
#Universal Transverse Mercator (UTM)
#       Cyl, Sph
#       zone= south
# +init=epsg:26711 +proj=utm +zone=11 +ellps=clrk66 +datum=NAD27 +units=m
# +no_defs +nadgrids=conus,ntv1_can.dat
#--- following specified but NOT used
# +ellps=clrk66
# ---- To Coordinate System ----
#Lat/long (Geodetic)
#
# +proj=latlong +datum=NAD27 +ellps=clrk66 +nadgrids=conus,ntv1_can.dat
The proj/nad/epsg file can be browsed and searched in a text editor for coordinate systems. There are known to be problems with some coordinate systems, and any coordinate systems with odd axes, a non-greenwich prime meridian or other quirkyness are unlikely to work properly. Caveat Emptor!

How do I use 3 parameter and 7 parameter datum shifting

Datum shifts can be approximated with 3 and 7 parameter transformations. Their use is more fully described in the towgs84 discussions.

Does PROJ.4 work in different international numeric locales?

No. PROJ.4 makes extensive use of sprintf() and atof() internally to translate numeric values. If a locale is in effect that modifies formatting of numbers, altering the role of commas and periods in numbers, then PROJ.4 will not work. This problem is common in some European locales.

On unix-like platforms, this problem can be avoided by forcing the use of the default numeric locale by setting the LC_NUMERIC environment variable to C.

eg.

$ export LC_NUMERIC=C
$ proj ...

Changing Ellipsoid / Why can't I convert from WGS84 to Virtual Earth Mercator?

The coordinate system definition for Virtual Earth Mercator is as follows, which uses a sphere as the earth model for the mercator projection.
+proj=merc +a=6378137 +b=6378137 +lat_ts=0.0 +lon_0=0.0 
     +x_0=0.0 +y_0=0 +k=1.0 +units=m +no_defs
But, if you do something like:
cs2cs +proj=latlong +datum=WGS84 
    +to +proj=merc +a=6378137 +b=6378137 +lat_ts=0.0 +lon_0=0.0 
                   +x_0=0.0 +y_0=0 +k=1.0 +units=m +no_defs
to convert between WGS84 and mercator on the sphere there will be substantial shifts in the Y mercator coordinates. This is because internally cs2cs is having to adjust the lat/long coordinates from being on the sphere to being on the WGS84 datum which has a quite differently shaped ellipsoid.

In this case, and many other cases using spherical projections, the desired approach is to actually treat the lat/long locations on the sphere as if they were on WGS84 without any adjustments when using them for converting to other coordinate systems. The solution is to "trick" PROJ.4 into applying no change to the lat/long values when going to (and through) WGS84. This can be accomplished by asking PROJ to use a null grid shift file for switching from your spherical lat/long coordinates to WGS84.

cs2cs +proj=latlong +datum=WGS84 
    +to +proj=merc +a=6378137 +b=6378137 +lat_ts=0.0 +lon_0=0.0 
                   +x_0=0.0 +y_0=0 +k=1.0 +units=m +nadgrids=@null +no_defs
Note the strategic addition of +nadgrids=@null to the spherical projection definition.

Similar issues apply with many other datasets distributed with projections based on a spherical earth model - such as many NASA datasets, and also (I think) the Google Maps mercator projection.


Requests to add items to the frequently asked questions list can be entered in bugzilla.