If your data are globally distributed (as you said they are), then UTM
probably isn't the best choice.  A global equal-area projection would
be better, in that you can use the same parameters regardless of where
on the globe you're working.  I have a similar application, and use
Cylindrical Equal Area.  I use something like (with an old version of
PostGIS -- I think the function names all start with ST_ now):

select sum(area(transform(glacier_polys,32767)))  FROM glacier_polygons ...

Note that the EPSG code here is one I defined.  I think the current
version of PostGIS might have CEA projection defined in some proper
location.

Bruce

On Mon, Sep 8, 2008 at 7:36 PM, Robert Coup
<[EMAIL PROTECTED]> wrote:
> On Tue, Sep 9, 2008 at 3:14 AM, Nicolas Ribot <[EMAIL PROTECTED]> wrote:
>>
>> No for surfaces. You will have to project your data into a planar
>> projection system to do so (UTM for instance). This thread may be
>> useful in your case to find the correct UTM zone according to your
>> objects coordinates:
>
> Untrue, the maths is just a little harder! And you need to assume the globe
> is a sphere (ala distance_sphere()):
> If you're dealing with lat-lon "rectangles" you want something like this:
> http://mathforum.org/dr.math/faq/formulas/faq.sphere.html#spherecap
> If you're dealing with arbitrary polygons its a bit more complicated:
> http://groups.google.com/group/comp.infosystems.gis/browse_frm/thread/af0401d07c3068b5?lnk=igtc
> HTH,
> Rob :)
> _______________________________________________
> postgis-users mailing list
> [email protected]
> http://postgis.refractions.net/mailman/listinfo/postgis-users
>
>



-- 
Bruce Raup
http://cires.colorado.edu/~braup/
_______________________________________________
postgis-users mailing list
[email protected]
http://postgis.refractions.net/mailman/listinfo/postgis-users

Reply via email to