XBeach : use of Swan specta
Laurent Leballeur
I would like to use Swan time-varying spectra as wave boundary
conditions of XBeach.
I'm not focusing on long wave contribution/influence within the
simulation results. I did Swan spectra outputs every 15 minutes during 6
hours. In my params.txt file, I specified rt=21600s and dtbc=900s. But
maximum recommended values for these 2 parameters are respectively 7200s
and 1s... It seems that I misunderstood something to well use this
functonnality.
Do I have to
Laurent Leballeur
Sorry for previous incomplete email...
Do I have to write a FILELIST file with several separate Swan spectra
every 15 minutes as the exemple in the documentation :
FILELIST
900 0.2 swan1.sp2
900 0.2 swan2.sp2
900 0.2 swan3.sp2
900 0.2 swan4.sp2
....
My log file at the beginning of the simulation points out the following :
"Changing dt in wave boundary conditions to satisfy Nyquist condition.
New dt = 0.5000 s."
What value do I set to dtbc ? Why should this value be lower than 1 ? I
do not want to do Swan spectral outputs every second, so what is the
solution ?
Best regards,
Laurent Leballeur
Robert McCall
For your situation you will have to generate a FILELIST file with 24
separate spectrum files (one for each 15 minute period). The first
column in this file (after the FILELIST header) is the value of "rt",
or the duration this wave spectrum applies (in your case rt = 15*60 =
900s). The second column is the value of "dtbc", which is an internal
parameter for XBeach. The value of "dtbc" is used to create the
internal time series of wave energy and mass flux at the model
boundary. The value should be small enough to reasonably describe the
short wave signal in time (~0.5-2 seconds). Generally this parameter
will not have a very large impact on your results, as long as a value
is selected within this range. So in short, "rt" controls how often a
SWAN spectrum is read, "dtbc" controls the generation of a random time
series from this spectrum. The total duration of your simulation (6
hours = 21600s) is set by the parameter "tstop" in params.txt.
If your SWAN output if all bunched in one .sp2 file, you will have to
break down this file into 24 separate files, i.e. one file per
spectrum, each with the correct SWAN header, including the frequency
and direction bins. You could take the stationary .sp2 file example
from the SWAN manual as a base, and then copy in the correct number
and values of angle and frequency bins, the FACTOR value and the 2D
spectrum, for each spectrum from your own SWAN output. Otherwise I
would recommend using the open-source tools on OpenEarth (http://
public.deltares.nl/display/OET/OpenEarth), specifically the XBeach and
SWAN tools written in Matlab (https://repos.deltares.nl/repos/
OpenEarthTools/trunk/matlab/applications/xbeach/xb_nesting/xb_swan/).
If you do not yet have an OpenEarth account, you will have to register
for one.
Robert