[GRASS-user] polar stereographic projection: strange behaviour of r.contour and d.grid

Hello!

I'm working on a map in polar stereographic projection and encountered two
problems:

r.contour
The module does not work in this projection.
I use the four northernmost tiles of the Globe DEM dataset. In the
original lat/long location r.contour works fine and fast. On the
reprojected tiles (with coarser resolution, 3000 m - ca. 10 % of the size
of the original tiles) r.contour starts as usual, but then harddisk
activity becomes very high and makes the whole system unresponsive for a
few minutes. The system becomes responsive again after r.contour exits
without result or message.
For me it works fine to reproject the contours from the lat/long location
but it shows that there is some kind of problem with r.contour.

d.grid
d.grid and its equivalent in the gis-manager do not produce acceptable
results in the polar stereographic location. The grid shows always an
irregular hole centred around the north pole, partly covering the larger
part of the displayed region. This hole is present in all zoom-levels.

Does anyone know about a way to construct the grid as a vector layer with
"round parallels" - all my attempts ended up with edges at the crossing
points with the meridians and straight lines inbetween.

Cheers,

Henning

Henning Lorenz wrote:

I'm working on a map in polar stereographic projection and encountered two
problems:

<snip>

Wolfgang,

What GRASS version and platform? Compiled yourself or prebuilt (the
source, if so)?

Does anyone know about a way to construct the grid as a vector layer with
"round parallels" - all my attempts ended up with edges at the crossing
points with the meridians and straight lines inbetween.

Maybe create the grid with v.mkgrid in a latlong location and then
reproject it to your stereographic location?

Another wild guess:

Create the center meridian 0 line, create as many meridians as needed
rotating it with v.transform. Create a vector point on the pole, create
parallels with v.parallel or v.buffer.

Maciek

Maciej Sieczka wrote:

Another wild guess:

Create the center meridian 0 line, create as many meridians as needed
rotating it with v.transform. Create a vector point on the pole, create
parallels with v.parallel or v.buffer.

... and merge them all with v.patch, forgot to add.

Maciek

What GRASS version and platform? Compiled yourself or prebuilt (the
source, if so)?

Sorry, I forgot to mention this. It's Grass6.3-cvs on Ubuntu 6.10. Grass,
proj and gdal are compiled on my system and are otherwise working fine.

Henning