[GRASS-user] Newbie question on g.region

[Michael Barton]

I don't know. In a remote sensing class I just co-taught with a
geomorphologist colleague, we did some comparisons between the NED and SRTM
data. The SRTM data were better at 30m at least.

Do a slope map (r.slope.aspect) of the NED data and see if it is striped,
indicating the contour line stepping.

Also subtract it from an SRTM to see if there is systematic error derived
from poor interpolation.

Michael
__________________________________________
Michael Barton, Professor of Anthropology
School of Human Evolution & Social Change
Center for Social Dynamics and Complexity
Arizona State University

phone: 480-965-6213
fax: 480-965-7671
www: http://www.public.asu.edu/~cmbarton


> From: gary <lists at lazygranch.com>
> Date: Fri, 09 Jun 2006 09:55:53 -0700
> To: Michael Barton <michael.barton at asu.edu>
> Cc: <grassuser at grass.itc.it>
> Subject: Re: [GRASS-user] Newbie question on g.region
> 
> http://seamless.usgs.gov/website/seamless/products/srtm1arc.asp
> 
> The link above indicates the 30m SRTMs are 1 arcsecond resolution. While
> I am going to first use the 1 arcsecond NED, there are 1/3 arcsecond
> NEDs available. Would the 30m  SRTMs be superior to a 1/3 arcsecond NED?
> 
> Michael Barton wrote:
>> The new GRASS 6.1 startup has an option to create a location from a
>> georeferenced map.
>> 
>> If you already have a location (i.e., projection) set up and something in
>> it, but are having region issues, you can use g.region to set the region of
>> interest to match the map you have
>> 
>> G.region rast=[mapname]
>> 
>> This will set up the region properly to match the resolution and extents of
>> your map.
>> 
>> I should also point out that NED are derived from topo contour lines and
>> often have poorly interpolated topography, such that they have a step-like
>> quality. I'd recommend using the 30m SRTM's instead. I haven't had trouble
>> with r.los, but r.cva is a much more versatile tool. It's only drawback is
>> it's EXTREME slowness when doing multiple viewsheds.
>> 
>> Michael
>> __________________________________________
>> Michael Barton, Professor of Anthropology
>> School of Human Evolution & Social Change
>> Center for Social Dynamics & Complexity
>> Arizona State University
>> 
>> phone: 480-965-6213
>> fax: 480-965-7671
>> www: http://www.public.asu.edu/~cmbarton
>> 
>> 
>> 
>> 
>>> From: Hamish <hamish_nospam at yahoo.com>
>>> Date: Fri, 9 Jun 2006 21:24:47 +1200
>>> To: gary <lists at lazygranch.com>
>>> Cc: <grassuser at grass.itc.it>
>>> Subject: Re: [GRASS-user] Newbie question on g.region
>>> 
>>> 
>>>> My ultimate goal for GRASS is to do viewshed analysis with r.los. I
>>>> have downloaded NED files from the USGS seamless website. I have also
>>>> managed to convert the .adf file to geotiff using gdal_translate.
>>> 
>>> Can't you download NED data directly as GeoTIFF?
>>> 
>>> 
>>> Some hints:
>>> 
>>> r.los doesn't scale well at all. Download "r.cva" (install with GRASS GEM?)
>>> and use that instead if you can.
>>> 
>>> Also see hints on r.sun and r.sunmask page.
>>> 
>>> 
>>>> My problem concerns starting up GRASS. When I try to run r.in.gdal, I
>>>> get an error message regarding the default region is not set. I gather
>>>> I'm not initializing GRASS properly.
>>> 
>>> g.setproj ?
>>> 
>>> g.region -p ?
>>> 
>>> r.in.gdal -o ?
>>> 
>>> 
>>> 
>>> good luck,
>>> Hamish
>>> 
>>> 
>> 
>> 
>> 
>>