[Proj] Representing rotated poles

[Pascoe, S (Stephen)]

Hi,
 
I deal with climate model data and have regional model data defined on a
long/lat grid with the poles rotated away from their true position (in
order to keep the pole singularities away from the region of interest).
Typically this grid would be defined as follows:
 
grid_north_pole_latitude:  39.2 deg (true latitude of rotated north
pole)
grid_north_pole_longitude: 198.0 deg (true longitude of rotated north
pole)
north_pole_grid_longitude: 0 deg (longitude of the prime meridian in the
rotated coordinate system)
 
How would I represent this coordinate system with proj4 parameters?  I
have been trying to use +towgs84:
 
+proj=longlat
+towgs84=0,0,0,-grid_north_pole_longitude*60*60,90-grid_north_pole_latit
ude*60*60,north_pole_grid_longitude,0
 
So far this doesn't seem to work (I'm using gdal to transform an image
to EPSG:4326).  Am I on the right lines?  Standard GIS terminology such
as datum, spheroid etc. are unfortunately not part of the standard
vocabulary of climate modellers.
 
Many thanks,
Stephen.
 
---
Stephen Pascoe  +44 (0)1235 445980
British Atmospheric Data Centre
Rutherford Appleton Laboratory
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/proj/attachments/20080121/52599b31/attachment.html>