How to output energy density spectrum in ROMS/SWAN system

Message

chenyf · #1 Unread post by **chenyf** » Thu Apr 24, 2008 2:30 am

Hey all,

Im now using coupled system of ROMS and SWAN and could successfully get the result of Wave Height, Wave Direction. But i have problem with getting energy density spectrum.

Im just following the SWAN manual and add following into SWAN INPUT:
********************************************************
POINTS 'LITE' -75 37.5
SPEC 'LITE' SPEC1D 'spec.dat' OUTPUT 20050517.0000 1 HR
********************************************************

But it didnt work.

Actually there isnt any Error file output. But following error was reported:

+SWAN is processing output request 1
p1_12504: p4_error: interrupt SIGSEGV: 11
rm_1_1_12513: (0.695312) net_send: could not write to fd=5, error =32

And i was confused that if I didnt write for a energy spectrum output but just output for wave heigh, direction as well as other info, such as DEPTH, DISSIP, UBOT, wind, XP, YP,etc, it could work successfully.

Is there any difference between ROMS/SWAN coupled system and SWAN only system in getting the energy density spectrum???

Anybody who has successfully get the output for energy spectrum by using coupled system of ROMS/SWAN? Or get any idea of such kind of problem?

Thanks:)

jcwarner · #2 Unread post by **jcwarner** » Thu Apr 24, 2008 3:14 am

we write out 2D spec while using the coupled system. Your command line looks ok. Seems like the error is saying that it is having trouble writing to the new file 'spec.dat' (?).
"could not write to fd=5, error =32"

Try running it with just 1 processor to roms and just 1 to swan. That may give a more useable swan error file.

chenyf · #3 Unread post by **chenyf** » Thu Apr 24, 2008 3:24 am

Thanks a lot.

Actually i already tried doing just 1 for ROMS and 1 for SWAN and also I tried for SPEC2D but it still didnt work and without Error file.

Im really confused coz it failed just because of adding those two lines in the SWAN INPUT.

btw, in PRINT file, it just ends up with

COMPUTE NONSTAT 20050516.000000 240 SEC 20050530.000000

and without any hints about the error....

jcwarner · #4 Unread post by **jcwarner** » Thu Apr 24, 2008 12:40 pm

is this with the newest version?
What else are you send to swan output?
What if you move the call for the spec out to be the last output?

chenyf · #5 Unread post by **chenyf** » Thu Apr 24, 2008 1:34 pm

Im not using the newest version and Im using the version which was last-changed at 2008-01-09(Revision: 141)-- I remember it's version 3.0.

As for the newest version (3.1), actually one of my labmate just updated for it several days ago but cant compile it successfully with the following error:

A array named 'OutField' in roms_export.f90 (line 91)was not be declared.

And I was using the spectrum output as the last output.

Thanks:)

arango · #6 Unread post by **arango** » Thu Apr 24, 2008 2:37 pm

This was already corrected. See at the bottom of ticket 149. You need to update.

chenyf · #7 Unread post by **chenyf** » Mon Apr 28, 2008 5:52 pm

I just updated to the newest version this morning but still failed in energy spectrum output with the same problem...

jcwarner · #8 Unread post by **jcwarner** » Mon Apr 28, 2008 6:36 pm

here is another method to try.
Go to the swan web site
http://vlm089.citg.tudelft.nl/swan/index.htm
download the software, install on your system,
and try the same INPUT file and see if it works. You will probably have to comment out the lines for the currents and the water level, as you may not have those input files.
If that works then it has to do with the coupling. If not then it may be something about your system.

arango · #9 Unread post by **arango** » Mon Apr 28, 2008 6:51 pm

In the new distributed ROMS version (revision 146 or higher), if MODEL_COUPLING is not activated for SWAN you get their original code for version 40.51.