r.slope.aspect

[Susan Stitt]

  >From grass-lists-owner at moon.cecer.army.mil Mon May 15 08:38:17 1995
  >From: GREEN WILLIAM G <wg1938 at csc.albany.edu>
  >Subject: Re: r.slope.aspect
  >Sender: grass-lists-owner at moon.cecer.army.mil
  >Reply-To: grassu-list at moon.cecer.army.mil
  >To: grassu-list at max.cecer.army.mil
  >Content-Type  >:   >text  >
  >X-Lines: 51
  >
  >
  >Philip,
  >
  >Thanks for the suggestion, but the problem isn't large flat areas.  It appea
rs
  >that when there is a steep slope, every other "line" of pixels is showing a
  >slope of 0.  In the area that I'm working (West Point, NY) the relief is pre
tty
  >dramatic so I know this can't be correct.  I will try your suggestion of usi
ng
  >s.surf.tps anyway and let you know if it works.
  >
  >Bill Green
  >
  >
Bill,

I will hazard a guess in this discussion.  This is just a hypothesis, but if yo
u
have a DEM with a given cell size of, say 30 meters, and you run r.slope.aspect
with your region set to a smaller cell size, say 10 meters, then the automatic
nearest neighbor resampling in GRASS would generate slopes of 0 on every other
or every third raster.  You might re-examine your region settings, if this is
a possibility.  GRASS will not automatically generate new elevation values
in your dem on the fly, it will use nearest neighbor resampling, thus your regi
on 
should be no finer than the DEM from which you are working.  However, you could
 
run some surface generation routines on your DEM to interpolate a surface of fi
ner 
resolution (such as s.surf.tps), but I would suggest care, for the surface you 
generate will have no more accuracy than your original surface.  It will look
prettier, but will not necessarily be a better representation of the actual 
terrain.  You also may be generating a false smoothing of the slope and aspects
.
If you are familiar with the actual terrain, you will be able to make a
judgement on this.

Susan Stitt
National Biological Service
Technology Transfer Center
sstitt at its.nbs.gov