[GRASS-user] help with r.gwflow

[Vishal Mehta]

Hi Soren,

Thanks much for your example on how to use the river parameters. i'll see
about switching to Grass 7.

If you should be interested in collaborating in this kind of complex
application, do let me know. i'll send you a conference paper i'm writing
soon. after that, if you find it interesting enough and have the time, we
could perhaps develop it together and i can put you down as co-author if
that works out..

thanks for your patience with my repeated emails.FYI  i have done
distributed surface hydrological modeling 10 years ago in GRASS, and
eventually i'd like to do combined surface-groundwater modeling in GRASS.
This is the first time i've attempted groundwater modeling in GRASS.

Vishal

2012/7/19 Vishal Mehta <vishalm1975 at gmail.com>:
> > Thanks Soren,
> > for your script.
> >
> > i'm using r.gwflow in v6.5svn, which does not have the budget parameter
> > option. do you know how i can get the 7 version to work in my 6.5
> version?
>
> Unfortunately the grass7 version of r.gwflow does not run in grass6.5
> and i have no plans o back port the budget feature.
> So maybe you would like to switch to grass7 to compute the budget?
>
> > also, i'd like to know how all th rest if it is set up - the river_bed
> > parameters etc. do you have complete documentation of it? i'm trying to
>
> You need three parameters to model the river:
> 1.) The water table of the river in [m], mostly derived from a river
> vector map and converted into a raster map
> 2.) The bed of the river in [m], mostly derived from the river water
> table: r.mapcalc "river_bed = river_table - 2"
> 3.) The leakage coefficient of the river bed in [1/s]
>
> River table, river bed and leakage are occupying the same pixels.
>
> Simple and stupid example:
>
> # River with a water table of 20m and the bed at 18m with a
> # leakage coefficient of 0.0001 1/s
>
> # First convert the stream network into a raster map
> v.to.rast input=streams output=streams type=line
>
> # Then compute the water table of the river at 20m
> # and the height of the bed that is 2m lower than
> # the water table of the river
> r.mapcalc "river_head = if(isnull(streams), null(), 20)"
> r.mapcalc "river_bed = river_head - 2"
>
> # The river leakage coefficient can be computed from the hydraulic
> conductivity
> # of the river bed (0.0001m/s) and the thickness of the river bed (1m)
> r.mapcalc "river_leakage = if(isnull(streams), null(), 0.0001/1)"
>
> > model the coupled water supply-extraction and groundwater system of a
> city,
> > so more complex real world examples of the use of r.gwflow would be
> really,
> > really useful. have you used it in a ocomplex setting perhaps for your
> > dissertation?
>
> The only documentation that is available is the r.gwflow manual page
> and my diploma thesis [1] (in German)
> that describes the mathematical details. I never utilized r.gwflow to
> compute large complex real world problems.
>
> Best regards
> Soeren
>
> [1]
> http://www.hydrogeologie.tu-berlin.de/fileadmin/fg66/_hydro/Diplomarbeiten/2007_Diplomarbeit_Soeren_Gebbert.pdf
>
> >
> > thanks again,
> > Vishal
> >
> >
> > On Wed, Jul 18, 2012 at 2:47 PM, Sören Gebbert
> > <soerengebbert at googlemail.com> wrote:
> >>
> >> Hi Vishal,
> >>
> >> 2012/7/18 Vishal Mehta <vishalm1975 at gmail.com>:
> >> > hi Soren,
> >> >
> >> > my constan head boundary conditions on the edges are causing water
> >> > tables to
> >> > build up everywhere else. so i estimated a constant flux 0f 0.0019
> m3/s
> >> > that
> >> > i want to apply for edge cells. this is what i am trying in order to
> >> > impose
> >> > a constant flux on the edges
> >> >
> >> > r.mapcalc "sink.init=if(row()==1 || row()==444 ||col()==1
> >> > ||col==477,-0.0019,null())
> >> >
> >> > #sink.init is the constant flux on edges
> >> > # my r.gwflow script runs at monthly time step- ihave a loop
> >> > $month-here's
> >> > the r.gwflow snippet
> >> >
> >> > r.gwflow --o -s solver=cg top=top bottom=bottom status=bc2 hc_x=k.1
> >> > hc_y=k.1
> >> > s=s.1 type=unconfined dt=2592000 error=0.05 phead=sim.$((prevmonth))
> >> > r=gwnaturalms.$((month)) q=sink.init output=sim.$((month))
> >> >
> >> > my boundary condition map (bc2) has contant head in strea pixels. all
> >> > else
> >> > are calculated.
> >> >
> >> > do you think the above implementation will correctly impose the
> constant
> >> > flux on the edges? it seems though that since this is not a boundary
> >> > condition, how can it ensure a constant flux at edges?
> >>
> >> Your script looks reasonable, except that the error term is much to
> large.
> >> You may need to use a smaller number like 10⁻7,
> >> otherwise your results my be wrong.
> >>
> >> I have attached a small r.gwflow example to show you how to estimate
> >> the constant flux at
> >> a western boundary using the budget computation. Since the flow will
> >> be specified
> >> as a source term, it will be constant the whole computational time.
> >>
> >> Maybe you can apply this method to estimate the boundary flux in your
> >> area?
> >>
> >> JFYI i am using r.gwflow of GRASS 7.
> >>
> >> Best regards
> >> Soeren
> >>
> >> >
> >> > Thanks for any help you can provide,
> >> > Vishal
> >> >
> >> > On Mon, Jul 16, 2012 at 1:46 PM, Vishal Mehta <vishalm1975 at gmail.com>
> >> > wrote:
> >> >>
> >> >> Thanks Soren,
> >> >> From you response, can you please tell me how to do the following two
> >> >> tasks (which i dont find in the online manual for r.gwflow): the
> >> >> remainder
> >> >> of your comments i have figured out.
> >> >>
> >> >> - How can i compute the flux in [m^3/s] for each cell with r.mapcalc?
> >> >> and
> >> >>
> >> >> - i have set the bc of the edges and stream cells at constant head
> for
> >> >> now- how can i get the budget raster maps you mention?
> >> >>
> >> >> thanks again,
> >> >> Vishal
> >> >>
> >> >>
> >> >> On Sat, Jul 14, 2012 at 12:05 PM, Sören Gebbert
> >> >> <soerengebbert at googlemail.com> wrote:
> >> >>>
> >> >>> Hi,
> >> >>> sorry for the delay.
> >> >>>
> >> >>> 2012/7/11 Vishal Mehta <vishalm1975 at gmail.com>:
> >> >>> > Thanks Soren,
> >> >>> > That explains some of the results i'm getting, with water piling
> up
> >> >>> > above
> >> >>> > the surface in the edges in low-lying areas.
> >> >>> >
> >> >>> > Can you please tell me how i can change that to constant flux or
> >> >>> > constant
> >> >>> > head? If constant flux, should that be in m/s units?
> >> >>>
> >> >>> Constant flux can currently only be defined using sources/sinks with
> >> >>> unit [m^3/s], that is option q.
> >> >>> But i can add two new options (fn, fe) that defines the flux in
> >> >>> northern or eastern direction using the unit [m/s]
> >> >>> that will be multiplied internally with the northern or eastern face
> >> >>> area of the cell?
> >> >>>
> >> >>> Otherwise you need to compute the flux in [m^3/s] for each cell with
> >> >>> r.mapcalc.
> >> >>>
> >> >>> >
> >> >>> > My problem though is that i dont know what a constant flux or head
> >> >>> > at
> >> >>> > the
> >> >>> > edges should be set to. For now the only bc i have put in there
> >> >>> > deliberately
> >> >>> > (beyond the default you mention) is that i have set constant head
> in
> >> >>> > stream
> >> >>> > pixels. I'll have to let flow through at the edges but i have no
> >> >>> > idea
> >> >>> > what
> >> >>>
> >> >>> You can use the river boundary condition to specify the flux in
> stream
> >> >>> pixel.
> >> >>>
> >> >>> > that flow should be. Are there some ways of setting the edge
> >> >>> > conditions
> >> >>> > such
> >> >>> > that the gw evolution in the central areas of interest are not
> >> >>> > highly
> >> >>> > influenced?
> >> >>>
> >> >>> You can set the boundary of interest to constant head pixel and
> >> >>> compute the flow throw the boundary pixel using the budget option.
> >> >>> The resulting budget raster map shows the flow from active cell into
> >> >>> sources, sinks and constant heads in [m^3/s].
> >> >>>
> >> >>> Best regards
> >> >>> Soeren
> >> >>
> >> >>
> >> >>
> >> >>
> >> >> --
> >> >> Vishal K. Mehta, PhD
> >> >> Scientist
> >> >> Stockholm Environment Institute - US
> >> >> 133 D St Suite F
> >> >> Davis CA 95616
> >> >> www.sei-us.org
> >> >
> >> >
> >> >
> >> >
> >> > --
> >> > Vishal K. Mehta, PhD
> >> > Scientist
> >> > Stockholm Environment Institute - US
> >> > 133 D St Suite F
> >> > Davis CA 95616
> >> > www.sei-us.org
> >
> >
> >
> >
> > --
> > Vishal K. Mehta, PhD
> > Scientist
> > Stockholm Environment Institute - US
> > 133 D St Suite F
> > Davis CA 95616
> > www.sei-us.org
>



-- 
Vishal K. Mehta, PhD
Scientist
Stockholm Environment Institute - US
133 D St Suite F
Davis CA 95616
www.sei-us.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/grass-user/attachments/20120719/93480ae0/attachment.html>