[Geotiff] :Mercator_2sp

[Frank Warmerdam]

Russell Ron wrote:
> Good Morning,
>  
> Can anyone give any guidance as to the keys that should be used when specifying the Mercator_2sp projection? What we used to do, and was apparently accepted by ArcGis 9.2 no longer works with ArcGis 9.3. It seems impossible to satisfy all software!
>  
> My current 'best attempt' is (for example)
>  
>       ProjCoordTransGeoKey (Short,1): CT_Mercator
>       ProjLinearUnitsGeoKey (Short,1): Linear_Meter
>       ProjNatOriginLongGeoKey (Double,1): xxxx                
>       ProjNatOriginLatGeoKey (Double,1): yyyy                
>       ProjFalseEastingGeoKey (Double,1): 0                
>       ProjFalseNorthingGeoKey (Double,1): 0                
>       ProjScaleAtNatOriginGeoKey (Double,1): 1                
>       ProjStdParallel1GeoKey (Double,1): yyyy
>  
> Where yyyy is what I would call 'Latitude of true scale'. 
>  
> This does not seem a true definition of Mercator where I think the Latitude origin is always the equator!
>  
> However these keys seem to do what we want with most software that we use (but not GDAL!)

Ron,

There is only one Mercator formulation in GDAL (and in the libgeotiff
"normalize" code), and it assumes it is provided a scale at the origin
which is normally on the equator.

If you have a standard parallel at which the scale is 1, I guess it is
expected that you will somehow use it to compute the scale at the origin
and provide that instead.

The GDAL/libgeotiff handling of Mercator is pretty much defined by

   http://www.remotesensing.org/geotiff/proj_list/mercator_1sp.html

I must confess I don't know a good way of computing the scale at the
equator if given a latitude of true scale.  I suppose you could establish
two points at that latitude that are truly one meter apart (using some
non-mercator projection with this as a latitude of true scale) and then
see how far those points are apart in mercator to establish the scaling
needed to return it to a distance of one meter in mercator.

I would be open to adding proper support for Mercator_2SP if someone
can provide me with some guidance.  It is annoying that there is only
one CT_Mercator in libgeotiff, so it is going to be very messy to
normalize this projection in libgeotiff into a single set of parameters.

A ticket would be helpful to track this.

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