[GRASSLIST:2378] The g3.region settings ....

GRASS GRASS user
1

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I'm trying to get s.vol.rst to work for the computation of rainfall map
using a 50m resolution cell (1000x1000 cells grid).
Now, the site file that's input for s.vol.rst has 4 columns
x|y|z|%rainfall(mm/year)

When I tried to launch s.vol.rst the first time it told me that I had to
set the 3D region. SO I launched g3.createregion which reead the
defaults from the current 2D region. I gave him the following options:

g3.createregion t=2500 b=0 dres=1

as to resemble:

t=2500 --> max altitude in meters that I have on my dem;
b=0 --> min altitude on my dem ;
dres=1 -->vertical resolution on 1 meter.

Now, I used the maskmap option for s.vol.rst so as to narrow the
computation only to a subarea of the 1000x1000 cell area while all
rainfall station (25) fall in the 1000x1000 cell area.

I launched the program and it is takeing a hell of a long time to
compute. After 4 hours it's still at 0%.

Did I miss something in the settings of the 3D region or does it simply
take so much time to compute? Yet, 1000x1000 cells don't seem to be much
......

Thanks

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFAF62lIRaQQEznKxURAitcAKDPGvkoo9M/8eFYDt5XdU6XcpgIgQCfVT9a
w9G5ENU5S9tqCVKuzSjytMI=
=2iRF
-----END PGP SIGNATURE-----

2

Geotronix wrote:

I'm trying to get s.vol.rst to work for the computation of rainfall map
using a 50m resolution cell (1000x1000 cells grid).
Now, the site file that's input for s.vol.rst has 4 columns
x|y|z|%rainfall(mm/year)

When I tried to launch s.vol.rst the first time it told me that I had to
set the 3D region. SO I launched g3.createregion which reead the
defaults from the current 2D region. I gave him the following options:

g3.createregion t=2500 b=0 dres=1

as to resemble:

t=2500 --> max altitude in meters that I have on my dem;
b=0 --> min altitude on my dem ;
dres=1 -->vertical resolution on 1 meter.

Now, I used the maskmap option for s.vol.rst so as to narrow the
computation only to a subarea of the 1000x1000 cell area while all
rainfall station (25) fall in the 1000x1000 cell area.

I launched the program and it is takeing a hell of a long time to
compute. After 4 hours it's still at 0%.

Did I miss something in the settings of the 3D region or does it simply
take so much time to compute? Yet, 1000x1000 cells don't seem to be much

It's a volume, not an area: 1000x1000x2500 = 2.5x10^9 cells.

--
Glynn Clements <glynn.clements@virgin.net>

3

On Wed, Jan 28, 2004 at 01:40:06PM +0100, Geotronix wrote:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I'm trying to get s.vol.rst to work for the computation of rainfall map
using a 50m resolution cell (1000x1000 cells grid).
Now, the site file that's input for s.vol.rst has 4 columns
x|y|z|%rainfall(mm/year)

When I tried to launch s.vol.rst the first time it told me that I had to
set the 3D region. SO I launched g3.createregion which reead the
defaults from the current 2D region. I gave him the following options:

g3.createregion t=2500 b=0 dres=1

as to resemble:

t=2500 --> max altitude in meters that I have on my dem;
b=0 --> min altitude on my dem ;
dres=1 -->vertical resolution on 1 meter.

It's probably much better, if you eventually just want to extract a 2D
rainfall map, to run the computation in a single-layer volume:

g3.createwind t=2500 b=0 dres=2500

t=2500 --> max altitude in meters that I have on my dem;
b=0 --> min altitude on my dem ;
dres=2500 -->vertical resolution 2500 meter.

Then s.vol.rst ...

Please report if this is faster.

Markus Neteler

4

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

* mercoledì 28 gennaio 2004, alle 15:43, Glynn Clements scrive:

Geotronix wrote:

> I'm trying to get s.vol.rst to work for the computation of rainfall map
> using a 50m resolution cell (1000x1000 cells grid).
> Now, the site file that's input for s.vol.rst has 4 columns
> x|y|z|%rainfall(mm/year)
>
> When I tried to launch s.vol.rst the first time it told me that I had to
> set the 3D region. SO I launched g3.createregion which reead the
> defaults from the current 2D region. I gave him the following options:
>
> g3.createregion t=2500 b=0 dres=1
>
> as to resemble:
>
> t=2500 --> max altitude in meters that I have on my dem;
> b=0 --> min altitude on my dem ;
> dres=1 -->vertical resolution on 1 meter.
>
> Now, I used the maskmap option for s.vol.rst so as to narrow the
> computation only to a subarea of the 1000x1000 cell area while all
> rainfall station (25) fall in the 1000x1000 cell area.
>
> I launched the program and it is takeing a hell of a long time to
> compute. After 4 hours it's still at 0%.
>
> Did I miss something in the settings of the 3D region or does it simply
> take so much time to compute? Yet, 1000x1000 cells don't seem to be much

It's a volume, not an area: 1000x1000x2500 = 2.5x10^9 cells.

Do you mean that 2500 I set is in cells? I thought it was in meters.....
The 1000x1000 is a grid with 50 meter resolution. So to set z=2500 must
I give it 2500/50=50? So that I get 1000x1000x50?
Or maybe the vertical scale resolution is indipendant from the x,y ones?

Cheers
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFAGTdMIRaQQEznKxURAhRrAKCD3eyyP34TLBvf71Rcgsis/lylTQCg0DGf
Dy/0a76lLHsvyoFZc93HxK4=
=Cd3m
-----END PGP SIGNATURE-----

5

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

* giovedì 29 gennaio 2004, alle 10:44, Markus Neteler scrive:

On Wed, Jan 28, 2004 at 01:40:06PM +0100, Geotronix wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hi,
>
> I'm trying to get s.vol.rst to work for the computation of rainfall map
> using a 50m resolution cell (1000x1000 cells grid).
> Now, the site file that's input for s.vol.rst has 4 columns
> x|y|z|%rainfall(mm/year)
>
> When I tried to launch s.vol.rst the first time it told me that I had to
> set the 3D region. SO I launched g3.createregion which reead the
> defaults from the current 2D region. I gave him the following options:
>
> g3.createregion t=2500 b=0 dres=1
>
> as to resemble:
>
> t=2500 --> max altitude in meters that I have on my dem;
> b=0 --> min altitude on my dem ;
> dres=1 -->vertical resolution on 1 meter.

It's probably much better, if you eventually just want to extract a 2D
rainfall map, to run the computation in a single-layer volume:

g3.createwind t=2500 b=0 dres=2500

t=2500 --> max altitude in meters that I have on my dem;
b=0 --> min altitude on my dem ;
dres=2500 -->vertical resolution 2500 meter.

Then s.vol.rst ...

Please report if this is faster.

Markus Neteler

You're right! It's a lot faster.... but what does it change pratically
form using a dres=2500 instead of dres=1? IS the result diffrent?

There a re a few things that are not stated in the docs of s.vol.rst and
r.surf.rst.
What unit measure are used when I state smooth=1; tension=50 and
zmult=50. Are these percent values. If I don't know what units I'm
dealing with,how can I state some sort of mathimatical relation between
the values of these parameters and the geological/meteorological problem
I'm dealing with.
Let me make an example. Of my 25 stations, most of them have low
variance of the mean while a few have a very high variance. This lets me
think that I could use the smooth parameter set to zero for all the low
variance rainfall stations while I should use a higher smoothing for
the stations that have I higher variance. Is this conceptually correct?
But how do I relate the variance to the smoothing parameter?

Thanks
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFAGTnyIRaQQEznKxURAnDLAJ9D/vlZeevqHc6F5bpvoIWXcnbvAQCfcf+h
yrOeoxQliLNsz46xFRA7520=
=8c/e
-----END PGP SIGNATURE-----

6

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

* giovedì 29 gennaio 2004, alle 10:44, Markus Neteler scrive:

On Wed, Jan 28, 2004 at 01:40:06PM +0100, Geotronix wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hi,
>
> I'm trying to get s.vol.rst to work for the computation of rainfall map
> using a 50m resolution cell (1000x1000 cells grid).
> Now, the site file that's input for s.vol.rst has 4 columns
> x|y|z|%rainfall(mm/year)
>
> When I tried to launch s.vol.rst the first time it told me that I had to
> set the 3D region. SO I launched g3.createregion which reead the
> defaults from the current 2D region. I gave him the following options:
>
> g3.createregion t=2500 b=0 dres=1
>
> as to resemble:
>
> t=2500 --> max altitude in meters that I have on my dem;
> b=0 --> min altitude on my dem ;
> dres=1 -->vertical resolution on 1 meter.

It's probably much better, if you eventually just want to extract a 2D
rainfall map, to run the computation in a single-layer volume:

g3.createwind t=2500 b=0 dres=2500

t=2500 --> max altitude in meters that I have on my dem;
b=0 --> min altitude on my dem ;
dres=2500 -->vertical resolution 2500 meter.

Then s.vol.rst ...

Please report if this is faster.

Markus Neteler

BTW.

I just tried some new settings with s.vol.rst and I'm getting every time
more puzzled....

He are two tries I gave to s.vol.rst using a 50m resolution DEM, with
area 1000x1000 cells, as "cellin" to compute my rainfall data using 25
mean annual rainfall values for 25 rainfall stations:

a) tension=50 smoothing=0 zmult=1 gave me a map that did not resemble
any altimetric influence. The mean thickness of the rainfall water
table, spread over the 74Km^2 area (after it was masked down to from
original 1600Km^2 using the "maskmap" option) is 1109mm

b) tension=50 smoothing=0 zmult=50 gave me a complete diffrent map where
you can see that topografic influence on the rainfall. To my surprise
the mean water table was 1110mm!! Only 1mm difference for two complete
diffrent maps.

The problem is that I'm not understanding how zmult influences the
application of the altimetric information to the interpolating
algorithmn.

Another thing is: My highest altimetric rainfall station is at 900 meters
above sea level (which has a mean annual rainfall of 1250mm).
I have many mountainous areas that reach 1300 meters
a.s.l. and s.vol.rst is not computing any rainfall value higher than
1250mm/year. Shouldn't it be predicting higher values for higher
altitudes?

Thanks.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFAGUDsIRaQQEznKxURAtTMAKCS8jnqI7UqHdR0Xbe7hTqVsSnjwACeO4SR
Accq6Wtv0NZnoZEKEg9ROPU=
=4Kf+
-----END PGP SIGNATURE-----

7

Geotronix wrote:

> > I'm trying to get s.vol.rst to work for the computation of rainfall map
> > using a 50m resolution cell (1000x1000 cells grid).
> > Now, the site file that's input for s.vol.rst has 4 columns
> > x|y|z|%rainfall(mm/year)
> >
> > When I tried to launch s.vol.rst the first time it told me that I had to
> > set the 3D region. SO I launched g3.createregion which reead the
> > defaults from the current 2D region. I gave him the following options:
> >
> > g3.createregion t=2500 b=0 dres=1
> >
> > as to resemble:
> >
> > t=2500 --> max altitude in meters that I have on my dem;
> > b=0 --> min altitude on my dem ;
> > dres=1 -->vertical resolution on 1 meter.
> >
> > Now, I used the maskmap option for s.vol.rst so as to narrow the
> > computation only to a subarea of the 1000x1000 cell area while all
> > rainfall station (25) fall in the 1000x1000 cell area.
> >
> > I launched the program and it is takeing a hell of a long time to
> > compute. After 4 hours it's still at 0%.
> >
> > Did I miss something in the settings of the 3D region or does it simply
> > take so much time to compute? Yet, 1000x1000 cells don't seem to be much
>
> It's a volume, not an area: 1000x1000x2500 = 2.5x10^9 cells.

Do you mean that 2500 I set is in cells? I thought it was in meters.....
The 1000x1000 is a grid with 50 meter resolution. So to set z=2500 must
I give it 2500/50=50? So that I get 1000x1000x50?

Note:

> > g3.createregion t=2500 b=0 dres=1

> > t=2500 --> max altitude in meters that I have on my dem;

> > dres=1 -->vertical resolution on 1 meter.

2500 metres with 1-metre resolution => 2500 cells.

Or maybe the vertical scale resolution is indipendant from the x,y ones?

Yes; that is what dres=... is for.

--
Glynn Clements <glynn.clements@virgin.net>

8

On Thu, Jan 29, 2004 at 05:50:58PM +0100, Geotronix wrote:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

* giovedì 29 gennaio 2004, alle 10:44, Markus Neteler scrive:
> On Wed, Jan 28, 2004 at 01:40:06PM +0100, Geotronix wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> >
> > Hi,
> >
> > I'm trying to get s.vol.rst to work for the computation of rainfall map
> > using a 50m resolution cell (1000x1000 cells grid).
> > Now, the site file that's input for s.vol.rst has 4 columns
> > x|y|z|%rainfall(mm/year)
> >
> > When I tried to launch s.vol.rst the first time it told me that I had to
> > set the 3D region. SO I launched g3.createregion which reead the
> > defaults from the current 2D region. I gave him the following options:
> >
> > g3.createregion t=2500 b=0 dres=1
> >
> > as to resemble:
> >
> > t=2500 --> max altitude in meters that I have on my dem;
> > b=0 --> min altitude on my dem ;
> > dres=1 -->vertical resolution on 1 meter.
>
> It's probably much better, if you eventually just want to extract a 2D
> rainfall map, to run the computation in a single-layer volume:
>
> g3.createwind t=2500 b=0 dres=2500
>
> t=2500 --> max altitude in meters that I have on my dem;
> b=0 --> min altitude on my dem ;
> dres=2500 -->vertical resolution 2500 meter.
>
> Then s.vol.rst ...
>
> Please report if this is faster.
>
> Markus Neteler

You're right! It's a lot faster.... but what does it change pratically
form using a dres=2500 instead of dres=1? IS the result diffrent?

Note, I am not the author of s.vol.rst.

The author once explained to me that, when being interesting in a
2D output map, it is not necessary to calculate the temporal volume
in descrete layers. so I expect an identical result.

There a re a few things that are not stated in the docs of s.vol.rst and
r.surf.rst.
What unit measure are used when I state smooth=1; tension=50 and
zmult=50. Are these percent values. If I don't know what units I'm
dealing with,how can I state some sort of mathimatical relation between
the values of these parameters and the geological/meteorological problem
I'm dealing with.

I write in CC to the authors, they may have an answer for this.

Let me make an example. Of my 25 stations, most of them have low
variance of the mean while a few have a very high variance. This lets me
think that I could use the smooth parameter set to zero for all the low
variance rainfall stations while I should use a higher smoothing for
the stations that have I higher variance. Is this conceptually correct?
But how do I relate the variance to the smoothing parameter?

Thanks

Markus