Repost of response on R.WATER.FEA

[Baxter E. Vieux]

> We are planning to use Grass here in France for an Environmental project which includes flood forecasting using rainfall previsions and real time rain gage dataand river water level acquisition.
> 
> A watershed analysis must be performed first with DEM or TIN data.
> We plan to use the r.watershed function to do that.
> My first question is :
> Does anyone knows if there is an easy way to obtain watershed basin contours using Grass from DEM as r.watershed does seem to do it.

You can use r.contours to generate contours from a DEM. But this is not needed
to run r.water.fea or r.watershed.
> 
> My second question is :
> 
> We found in the incomings two packages that look interesting for our purpose :
> 
> The First one from B Vieux and N. Gaur is 
> r.water.fea 
> which models overland and channel flow in a watershed in an interactive way.
> This package is well documented and we could get source code so as to be able to compile it.
> So we would like to hear about some people who tried this software tool before us to get their overall impressions on it.

Since Nalneesh and I wrote r.water.fea we are partial to it. We have documented
and tested it extensively. For example, we have set up watersheds of simple
geometry such as plane surfaces, and compared the numerical simulation with
analytic solutions. Therefore, we know that the FE/FDM algorithms are stable
accurate and are relatively fast considering the large number of grid cells
composing the watershed basins.

> Does someone may tell us whether it is possible to have the model run non interactively?
> 
We only have an interactive version to date. When CERL contracted with us
to develop the "Definitive Distributed Hydrologic Model" we anticipated that
users would prefer interactive. Please note that once you have set up a 
project and performed initial steps of defining maps, you can rerun without
starting over from the beginning.

> The second one is 
> r.hydro.CASC2D  from B. Saghafian and F. L. Odgen.
> 
> We did not found any source code for this one, just binaries and a short MAN note.

I am curious too as to how this came out. They tried porting a PC version of
this model to unix about the same time we were writing our version. 
> Can anyone give us some hints on where to get the source code or get in touch with the authors?

You can get in touch with me at:

	vieux at chief.ecn.uoknor.edu 

Also, please check out the home page which describes our current research
in GIS/hydrology. Open the URL at:

	http://chief.ecn.uoknor.edu/

As you will notice on the home page I am teaching a GIS course and a Hydrology
course using MOSAIC. However, I have not opened access from outside our domain
for these items. In the GIS course we essentially cover GRASS and then build
data sets of soils, topography, NEXRAD rainfall for simulation using
r.water.fea. We recode the soils into Green and Ampt infiltration parameters,
process NEXRAD rainfall into a time series of rainfall rate maps, and then
calibrate the model for observed events on a 1200 sq. km. basin. This of
course with our own GRASS software we have developed at OU.

I am listing below some of our published literature relevant to r.water.fea
and the "science of distributed hydrologic modeling."

Regards--
Baxter Vieux

~r /u/vieux/WP/NSF-EPSCOR/ARI/SIG*


cheers--
Baxter