r.answers

[Chris Rewerts]

>>From: delklo at leda.orstom.fr (DELCLAUX Frank)
>>Newsgroups: info.grass.user
>>Subject: r.answers
>>Date: 3 Aug 94 07:19:09 GMT
>>Message-ID: <9408030719.AA07649 at leda.orstom.fr>
>>Reply-To: grassu-list at max.cecer.army.mil
>
>hello,
>I am trying to use r.answers. With the sample database, it is OK, but
>with mine, when choosing option "No subsurface tile drainage in the
>watershed" I get the message "The output file seems incomplete. Can't
>find Channel Deposition data which should be at the bottom of the
>file". The diagnostic error message is "Soil type 90 specified for
>element 120, 821 is not defined" I just have 1 soil  type and elements
>120 and 821 look OK.
> When choosing the two others options, "All of the area of the
>watershed/part of... has subsurface drainage", the program abort, it could
>not execute answers "Perhaps the program answers is not available".
>The input maps and files seems OK, even if I had some problems running
>r.direct and r.slope.
>Thank you very much in advance.
>
>Francois Delclaux.
>


Hmmmm. This sounds like an unlovely salad of warnings and errors...

>The output file seems incomplete. Can't
>find Channel Deposition data which should be at the bottom of the
>file

ok, keep in mind that r.answers uses GRASS maps and other inputs
you give it to create an input file for ANSWERS, then it runs ANSWERS
with that input file and collects the output from stderr and stdout.
After it runs ANSWERS it first shows you what it got from stderr, if
anything. Then it goes to the task of reading through the ANSWERS output
sent to stdout (which is saved in a file that you can access using
option 12 of the r.answers main menu). As it reads through the ANSWERS
output it is checking for things that it expects to see. If something
is missing, this probably means that something went wrong and ANSWERS
aborted the simulation. 

Thus, when it says it didn't find Channel Deposition at the bottom
of the file, we can infer that the simulation aborted, there must
be an error in the inputs...

>Soil type 90 specified for
>element 120, 821 is not defined" I just have 1 soil  type and elements
>120 and 821 look OK.

apparently when it read the soil map if found category value 90 for
the two elements named. put the watershed MASK into place, set the
region to that you are using for your simulation scenario, and look
at the map you are using for soils (or run r.report) to see if more
than category 1 shows up.

>Perhaps the program answers is not available

I kind of doubt this is the problem. At any rate, this warning is 
triggered if r.answers suspects that ANSWERS is either not compiled
(note that 'answers' is a fortran program and requires f77 or a similar
compiler) or it is not in the current path when r.answers runs...
A test would be to type 'which answers' in the GRASS command shell
to see if it says 'command not found' or if it returns the full
path to the answers binary. If it is not found you will need to 
run gmake in the answers src directory. This will be something like:
$GISBASE/src.garden/answers/src.answers/raster/r.answers/answers

Chris Rewerts                   Environmental Planning and Management Team
US Army Construction Engineering Research Laboratories, Champaign, IL, USA
rewerts at zorro.cecer.army.mil