[Proj] OSTN02 in NTv2 format for OSGB36<-->ETRS89 datum conversions

Paul Kelly paul-grass at stjohnspoint.co.uk
Mon Nov 13 07:04:03 EST 2006


Hello all
There was some interest in having the Ordnance Survey of Great Britain's 
"OSTN02" gridshift-based transformation from the OSGB36 to ETRS89 datums 
more widely available in a form that could be used with PROJ.4 (and thus 
also GRASS, GDAL etc).

I've managed to convert it using a combination of PROJ.4, GRASS, and Unix 
text processing tools. The process was so involved that I'm thinking of 
writing a short paper on it, but in the meantime I'll make the 
not-fully-tested or verified NTv2 files available for anyone who might 
want to use them. You can put either or both the files in your PROJ data 
directory, and use them in a co-ordinate system string like: 
+nadgrids=england-wales.gsb or +nadgrids=scotland.gsb,england-wales.gsb 
making sure there are no other +datum or +towgs84 parameters in the 
co-ordinate system specification. The +nadgrids line should be specified 
in the OSGB36 co-ordinate system, while the other one should have 
+towgs84=0,0,0.

Both the files have a resolution of 1 arc-minute of latitude and 40 
arc-seconds of longitude and are about 4 megabytes in size. This 
corresponds roughly to the 1 kilometre grid used in OSTN02, but I have a 
feeling it may be over the top and I'd like to do some tests to see if I 
can get a similar accuracy without the need for such a dense grid.

england-wales.gsb
Download from http://www.stjohnspoint.co.uk/gis/england-wales.gsb

Coverage:
N 56N
S 49d55'N
E 2E
W 6W

scotland.gsb
Download from http://www.stjohnspoint.co.uk/gis/scotland.gsb

Coverage:
N 61N
S 54d30'N
E  0d30'W
W  8W

All the latitiude/longitude values are referenced to the Airy ellipsoid.
http://www.stjohnspoint.co.uk/gis/total-coverage.png shows graphically the 
approximate coverage and overlap between the two files. The coastline data 
is from a very old dataset that used to be distributed with GRASS. I have 
no idea of the significance of the sharp change in latitude shift (i.e. the 
raster data plotted) beyond the coast. I expect the data beyond there is 
meaningless.

I can't give any guarantees as to the accuracy and usefulness of the 
files. In future, when I have it properly documented you can decide for 
yourself.

Best regards

Paul



More information about the Proj mailing list