open TELEMAC-MASCARET The mathematically superior suite of solvers

Good afternoon everybody,

I would ask you if it is possible with Telemac2D to take into account the process of infiltration.

Thank you in advance,

Daniele

Hello,

Yes. Given the fact that we deal with depth-averaged values, there is no difference between infiltration and evaporation, if you just want to take into account water that goes into the bottom. This would be like evaporation. It can be varying in space at the cost of implementing the variation into subroutine prosou.f. We used to have a connection with a groundwater module but this is not available at the time being.

With best regards,

Jean-Michel Hervouet

Thank you very much jmhervouet.

I would kindly ask you a suggestion/opinion about the following case:
I would like to use Telemac2D to simulate flooding scenarios by considering infiltration.
As you suggested, I'm gonna use the option "Rain or Evaporation" to take into account the part of water to remove.
My goal is to implement the variation of infiltration according to the soil types. I have a map with the soil types in a .xyz format where each pixel is represented by a CN value. Therefore, according to the code I would like to implement, infiltration will change according to the CN value of the pixel.
I was thinking to use Blue Kenue to add a CN variable to the Selafin file (right click on Selafin file, Add Variable, etc).
My question is the following: which are the subroutines I have to modify In order to tell Telemac to take into account this new variable?

Thank you in advance,

Daniele

Hello,

The subroutine where the rain becomes a space variable array is prosou.f (line 450). In this subroutine you can inspire from the calls to subroutine FIND_IN_SEL to see how to look for a variable with a given name in a selafin file. It is possible to put your variable in the geometry file with Blue Kenue, and to read it in prosou.f.

With best regards,

Jean-Michel Hervouet

Daniele and Jean-Michel,

I take the opportunity of this post to add a comment about a possible new feature to be 1implemented in the system and allowing to use user variables stored in the geometry file (JMH suggested that PRIVATE ARRAYS could be used to do the job). See the corresponding post here:
www.opentelemac.org/index.php/kunena/fee...s-from-geometry-file

Can this feature be implemented in a future release?

Otherwise, reading a variable from the geometry file can be done using FIND_IN_SEL from bief library. An example is available in some of the modified noerod.f subroutine that you might find in the Sisyphe part of the forum (I posted one in the past at least).

Good luck and best regards
PL

Hello,

Yes, it is an excellent suggestion and I'll add it in the next release. I propose a keyword USER VARIABLES that will be a block containing as many variables as wanted, with possibly a few pointers to the double precision arrays to avoid the parafernalia of USER_VARIABLES%ADR(2)%P%R(I), e.g.:

USER_VARIABLE1=>USER_VARIABLES%ADR(1)%P%R
USER_VARIABLE2=>USER_VARIABLES%ADR(2)%P%R

etc.

These variables will be like PRIVE arrays but will be read in the geometry file if present.

With best regards,

Jean-Michel Hervouet

Thank you Jean-Michel, these are great news!!

Best regards
PL

Dear Jean-Michel and PL,

I really appreciate your support, thank you very much again.
According to what I've understood so far, I modified the prosou.f file in the following way:
-) I added these lines in the declaration variables part of the subroutine:

CHARACTER(LEN=5) NAME
INTEGER :: t2dgeo

-) I added the following at line 395:

CALL BIEF_ALLVEC(1,CN,'CN',IELM1, 1 , 2 ,MESH)
NAME='CN'
CALL FIND_IN_SEL(CN,NAME,T2D_FILES(T2DGEO)%LU,
& T2D_FILES(T2DGEO)%FMT,W,OK,NPTH,NP,ATH)


So far, I should have told Telemac to read the CN input variable contained in the GEO file, isn't it?
I still have to implement the code for the infiltration according to CN number. Before continue I would like to know your opinion about the code I wrote.
I re-compiled Telemac and there are no error. So, I don't know how to understand if everything is going on well.

Thank you in advance,

Daniele.

Hello,

It looks about OK at first sight. The only problem is that the name of the variable should be given in 32 cahracters as it is in the serafin format (16 for the name, 16 for the unit). If I remember well only the first 16 characters are tested.

With best regards,

Jean-Michel Hervouet

Thank you again Jean-Michel Hervouet.
I did what you said and re-compile T2D.
Unfortunately when I try to run a simulation with the new variable I get this error:

...
BACK FROM CHARAC
CALLING PROSOU
LIT : ABNORMAL END OF FILE
ONE INTENDED TO READ
A RECORD OF 1 VALUES
OF TYPE : CH
ON LOGICAL UNIT : 14

PLANTE: PROGRAM STOPPED AFTER AN ERROR
...


I've checked in the forum to find a solution but with no luck.
I attach a copy of the prosou.f file. Modifications are after line 395 (before RAIN-EVAPORATION).

Any suggestion would be appreciated,

Daniele