Even though the projection is not defined, the data still has some information it is WGS84 which has a datum/coordinate system so you do need to define it, the EPSG code is 4326.
You can't reproject something from nothing, it needs at least a coordinate system, otherwise it doesn't know what point on your current map should be moved and to where it should be moved. When you configured your first location you should have specified the EPSG as 4326. Does that get you further? Alex Corrado wrote: > Dear Daniel, dear Nikos, dear list, > > I have tried everything I could. I read the GRASS books chapters and the > documentation. It does not work. > > The problem is that the files I have are in WGS84 unprojected. See here: > > http://www.worldclim.org/format.htm > > Whichever procedure I use, whther Daniel's or the book (for example: page > 47), > the problem is always the same. > > I create a location + mapset + map with the worldclim data (location01). I > load the original data, and I call the raster map "precipitation". I generate > the correct box with v.region and I call it "test". I create a new location + > mapset (utm + osgb), called location02. I run v.proj in location B pointing > at map "test" > > v.proj input=test location=location01 mapset=PERMANENT > > > I have the same problem: > > PROJ_INFO file not found for location location01 > > But of course the projection is not defined, the data are unprojected! > > I am slowly getting desperate .... I have been working on it several days and > cannot get a sensible result. I have tried with different datum > transformation parameters for osgb36 and trying to project .... but the > results are very unreliable .... > > Do you know what should I do? > > On Thursday 23 October 2008 11:38:47 Daniel Victoria wrote: >> Corrado, >> >> What v.in.region does is create a vector that surrounds your current >> region. Then, when you run v.proj, the vector covering the region is >> brought to your target location so you know where your projected >> raster should be at... >> >> The "manual labor" after that is setting the resolution correctly. You >> could then use some rule of thumb like 3arcsec ~= 90 m or, try to set >> the number of columns/rows in your target region similar to your >> lat/lon region. But this latter option does not work so well if your >> raster is to be tilted a lot... >> >> On an "additional" question: How do people set the resolution >> correctly when transforming from lat/lon to meter and vice-versa? Are >> there any other methods besides the ones I mentioned? >> >> Cheers >> Daniel >> >> On Thu, Oct 23, 2008 at 8:16 AM, Nikos Alexandris >> >> <[EMAIL PROTECTED]> wrote: >>> On Thu, 2008-10-23 at 10:40 +0100, Corrado wrote: >>>> Thanks Nikos, thanks Daniel! >>>> >>>> Unfortunately, my map is a raster map, not a vector .... can I still >>>> use >>>> v.in.region and v.proj? >>>> >>>> Thanks >>> [...] >>> >>> Hi Corrado! >>> >>> For sure!! You can use v.in.region anytime since it deals with the >>> current region and not with a specific map. >>> >>> Regards, Nikos > > > _______________________________________________ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user
