[GRASSLIST:4340] Re: GIS visitor requests directions

Glynn Clements glynn.clements at virgin.net
Tue Aug 20 16:48:11 EDT 2002


Nigel McFarlane wrote:

> ahh - I'm buried in terms again. I thought that "datum' related
> to Z-elevation only. I don't understand how arguments about different
> datums (data?) have any bearing on what 2D projections are
> supported. I don't know what NAD27 and NAD83 are.

NAD27 and NAD83 are datums. AFACIT, a datum determines the
correspondence between latitude/longitude/altitude and a specific
location.

> My question is: can GRASS take a set of data in one
> 2D projection (even if its just lat-long) and project that
> data into another 2D projection, regardless of which projections
> the two are, and regardless of what part of the earth the data
> covers?

Yes. Well, almost; both projections must be implemented in the PROJ
library which the *.proj programs use and, for one of them[1], the
inverse projection must also be implemented. The PROJ library
implements ~120 projections, including Plate Caree and UPS; for most
of those, the inverse is implemented.

[1] The source projection for [svm].proj, the destination projection
for r.proj.

> If so, I'm doing it wrong because I can't calculate
> Northing and Easting boundaries exactly to start with.

Note that you can change the region at any time with g.region; the
region which you specify when creating the location is only the
default.

A "brute force" option for determining the region is to convert the
source raster map to a sites map with r.to.sites, project the sites
map with s.proj, then set the region with "g.region sites=...". The
reason for using a sites map is that these aren't clipped to the
region when projecting, while rasters, by their nature, cover a fixed
region.

> Or, as this thread seems to indicate, are projections only
> supported on a case-by-case basis? (in which case, back to manifold)

I suspect that you're getting confused between datums and projections. 

The reason that datum conversions in *.proj are limited to
NAD27<->NAD83 is basically due to the fact that someone wanted
NAD27<->NAD83 conversion badly enough to add code to handle that
particular case.

Conversion between projections is "general case", in that the *.proj
programs will perform conversions between any combination of the 100+
"known" projections.

-- 
Glynn Clements <glynn.clements at virgin.net>



More information about the grass-user mailing list