[GRASSLIST:4326] Re: GIS visitor requests directions

[Glynn Clements]

Victor Wren wrote:

> > > I cannot re-project anything to another projection. I've read and
> > > re-read about setproj and numerous other commands. Can't do it.
> > > Don't know what the right steps are. No idea. I've messed around
> > > extensively and never got close. I've read just about everything
> > > on GRASS that google reports.
> > 
> > To reproject the data, first create a new location which uses the
> > desired target projection. Then, from within the new location, project
> > each map from the old location with r.proj, v.proj or s.proj.
> 
> As I understand this (based on the man pages and experimentation) 
> this ONLY works when projecting between NAD27 and NAD83.

Prior to pre5, the datum was ignored, so it would only work when
projecting between locations having the same datum.

In pre5, it should also handle conversions between NAD27 and NAD83,
but it always uses the "conus" conversion table.

> I have 
> attempted to use this to translate points between State Plane and 
> UTM coordinates (both NAD83), and the results are not good (i.e. the 
> translated vectors plot about 200 miles from the region of interest).  
> The documentation for ?.proj states clearly "Currently r.proj does not 
> support general datum conversions. It only supports conversion 
> between nad27 and nad83, and only within the CONUS conversion 
> grid; 20 degrees to 50 degrees north latitude, 63 degrees to 131 
> degrees west longitude. That covers all of the conterminous USA plus 
> Mexico north of Mexico City and most of Canada farther south than 
> Winnipeg, Manitoba." (notices for s.proj and v.proj are similar).  It 
> would also appear that the source and destination MUST be in the 
> same cross-section, i.e. State Plane or UTM.
> 
> I have a strong need for this, too, as I am dealing with several sources 
> of data, mostly in UTM (NAD27 and NAD83) State Plane (NAD83) and 
> Geo coordinates, but the customer requires State Plane (NAD83) for 
> the final product (and being in the USA, the data is a complete mix of 
> feet and meters).  Fortunately all of my data is within Colorado Central, 
> but ?.proj does not appear, currently, to be useful for world maps, or 
> for GEO coordinates of any kind.

One problem with coordinates which are in feet is that GRASS
interprets "foot" as international feet rather than US survey feet
(likewise for "inch"). You should be able to obtain US survey feet by
manually changing the "meters" entry in the PROJ_UNITS file from
0.3048 to 0.304800609601219.

However, I doubt that this will result in an error of 200 miles. 
Actually, I'm not sure what could result in an error of 200 miles.

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