[OSRS-PROJ] Datum transformation - more Sphere to WGS84...

[Frank Warmerdam]

Sverre Planke wrote:
> Hello GMT and PROJ.4 users:
> 
> Thank you for help and comments.
> 
> I need to convert the coordinates on a Spherical earth
> (GMT ELLIPSOID: "Sphere", 1980, 6371008.7714, 6371008.7714, 0.0)
> (e.g. Sandwell/Smith v.9.1 gravity data) to WGS84, but I
> have problems finding the right transformation parameters.
> 
> * I have installed the USGS software PROJ.4 from
>   http://www.remotesensing.org/proj/
>   Program cs2cs supports tranformations between
>   different datums if you have the right paramters (towgs84)
> 
> * A comprehensive list of "towgs84" paramters is found
>   on http://mapserver.gis.umn.edu/doc36/proj.html
>   However, not Sphere to WGS84 can not be fond on this site
>   or other places I have looked on the internet.
> 
> Question:
> Can anybody supply me with the right towgs84 paramters for
> a translation from Sphere to WGS84?

Sverre,

First, you are talking about converting from spherical coordinates.
In this case there is no real "datum" in effect, so you are just
converting fro spherical coodinates to ellpisoidal coordiantes on the
target ellipsoid (WGS84).  Currently PROJ only supports this sort of
transformation if you fake it into thinking you know the datum shift
required, but you can safely give a 0,0,0 shift.   I would do this like
this:

% cs2cs -v +proj=latlong +a=6371008.7714 +b=6371008.7714 +towgs84=0,0,0 +no_defs \
        +to +proj=latlong +datum=WGS84 << EOF
? -114 44
? 0 0
? 0 90
? EOF
# ---- From Coordinate System ----
#Lat/long (Geodetic)
#
# +proj=latlong +a=6371008.7714 +b=6371008.7714 +towgs84=0,0,0 +no_defs
# ---- To Coordinate System ----
#Lat/long (Geodetic)
#
# +proj=latlong +datum=WGS84 +ellps=WGS84 +towgs84=0,0,0
114dW   44d11'32.033"N 3217.845
0dE     0dN -7128.229
0dE     90dN 14256.457

While I have specified your particular spherical size, it doesn't really
matter if you only want 2D coordinates.   In this case it suggests that
converting from (114W, 44N) on your sphere would put you at (114W,44d11'32"N)
but 3218m below the surface of the sphere.  I have also converted an
equatoral and pole location to show the extreme effects on elevation at
those locations.

I would also add that the +no_defs on the spherical coordinate system
is very important in this case, otherwise somehow the +ellps=WGS84 default
from proj_def.dat will override the +b setting.

Best regards,

-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent


----------------------------------------
PROJ.4 Discussion List
See http://www.remotesensing.org/proj for subscription, unsubscription
and other information.