[GRASS-user] r.param.scale issue

[Vincent Bain]

Hi Markus,
thank you for your reply and investigation!

just in case, if I follow your first method, here's what it gives me on
a 12 steps loop, i.e. 40% success :

md5sum Bureau/slope_test.*
0bfb867317c149e9eb0b232ed6859294  Bureau/slope_test.1
872be645bdcc826442010e0f7cdcf8b3  Bureau/slope_test.10
4b5a343a333056c912728fcc7be0c314  Bureau/slope_test.11
1231556eb6c1307834d3eb0c41108982  Bureau/slope_test.12
1231556eb6c1307834d3eb0c41108982  Bureau/slope_test.2
bb89bd377a2c42f051dafe56318bff66  Bureau/slope_test.3
bb89bd377a2c42f051dafe56318bff66  Bureau/slope_test.4
1231556eb6c1307834d3eb0c41108982  Bureau/slope_test.5
1231556eb6c1307834d3eb0c41108982  Bureau/slope_test.6
1231556eb6c1307834d3eb0c41108982  Bureau/slope_test.7
4b5a343a333056c912728fcc7be0c314  Bureau/slope_test.8
4b5a343a333056c912728fcc7be0c314  Bureau/slope_test.9

my setup:
a freshly built GRASS GIS 7.9.dev (64 bit), code revision: 8c2c33b4f
Debian GNU/Linux 10.5 (buster)
the whole hardware is 4 months old...

I had the same issue with several raster source maps.

Well, for now I go back to r.slope.aspect...

V.


Le mercredi 14 octobre 2020 à 21:29 +0200, Markus Neteler a écrit :
> Hi Vincent
> 
> On Sun, Oct 11, 2020 at 8:08 PM Vincent Bain <bain at toraval.fr> wrote:
> > Dear Grass users,
> > has anyone ever experienced such issue with r.param.scale?
> > 
> > Here's the point:
> > I use to invoke r.param.scale instead of r.slope.aspect to
> > calculate
> > topographical slopes in various situations. I especially appreciate
> > to
> > have a hand on the size of processing window. I rencently had
> > trouble
> > with the result of the following command:
> > 
> > g.region rast=topo
> > r.param.scale input=topo output=slope_test size=5 method=slope --o
> > 
> > It seems to run fine, but when I inspect the result it /sometimes/
> > turns out to be really buggy (see attached .gif).
> > As I said above it's a very strange behaviour; the command run on
> > the
> > same input file (same region settings) fails , let's say, in 10 %
> > cases.
> 
> Weird indeed.
> 
> I tried locally (btw: which GRASS GIS version do you use?):
> 
> GRASS nc_spm_08_grass7/user1:~ > g.region raster=elev_lid792_1m
> for i in $(seq 1 12) ; do r.param.scale input=elev_lid792_1m
> output=slope_test size=5 method=slope --o ; r.univar -eg slope_test >
> slope_test.$i ; done
>  100%
>  100%
>  100%
>  100%
>  100%
>  100%
>  100%
>  100%
>  100%
>  100%
>  100%
>  100%
> GRASS nc_spm_08_grass7/user1:~ > md5sum slope_test.*
> b74294ec45c02ca6a86ac81380af5a73  slope_test.1
> b74294ec45c02ca6a86ac81380af5a73  slope_test.10
> b74294ec45c02ca6a86ac81380af5a73  slope_test.11
> b74294ec45c02ca6a86ac81380af5a73  slope_test.12
> b74294ec45c02ca6a86ac81380af5a73  slope_test.2
> b74294ec45c02ca6a86ac81380af5a73  slope_test.3
> b74294ec45c02ca6a86ac81380af5a73  slope_test.4
> b74294ec45c02ca6a86ac81380af5a73  slope_test.5
> b74294ec45c02ca6a86ac81380af5a73  slope_test.6
> b74294ec45c02ca6a86ac81380af5a73  slope_test.7
> b74294ec45c02ca6a86ac81380af5a73  slope_test.8
> b74294ec45c02ca6a86ac81380af5a73  slope_test.9
> 
> Even more tests:
> 
> GRASS nc_spm_08_grass7/user1:~ > for i in $(seq 1 111) ; do
> r.param.scale input=elev_lid792_1m output=slope_test size=5
> method=slope --o ; r.univar -eg slope_test > slope_test.$i ; done
>  100%
>  100%
>  100%
> ...
> 
> GRASS nc_spm_08_grass7/user1:~ > md5sum slope_test.*
> b74294ec45c02ca6a86ac81380af5a73  slope_test.1
> b74294ec45c02ca6a86ac81380af5a73  slope_test.10
> b74294ec45c02ca6a86ac81380af5a73  slope_test.100
> b74294ec45c02ca6a86ac81380af5a73  slope_test.101
> ...
> b74294ec45c02ca6a86ac81380af5a73  slope_test.98
> b74294ec45c02ca6a86ac81380af5a73  slope_test.99
> 
> Looks all fine...
> 
> Maybe check if the computer disk is still ok?
> 
> Best,
> Markus
>