open TELEMAC-MASCARET The mathematically superior suite of solvers

hello all
I am new to tomawac and wave modelling i want to ask about how to determine some physical keywords in the cas file. I want to simulate the waves for the North sea
1-first for the boundary conditions:
the keyword TYPE OF BOUNDARY DIRECTIONAL SPECTRUM if it is equal to 1, I will need some other keywords that confuse me which is as follow:
- boundary mean fetch value and initial mean fetch value
according to my knowledge the fetch length correspond to constant wind speed and direction that not change for a period of time but in my north sea case the wind is unsteady and inhomogeneous.
my exact question for this is
what is the difference between boundary mean fetch value and initial mean fetch value? And how can I determine them for the case of constant wind, and the other one for wind that change in space and time?
2- how to determine the directional spread 1 and directional spread 2 at the boundary?
did i need a wave time series at the boundaries to determine them?
3- the subroutine specini.f is implemented only for the whole mesh points NPOIN2. how it is used for the boundary conditions if I set the keyword TYPE OF BOUNDARY DIRECTIONAL SPECTRUM?

thanks in advance for your effort with me

Hello,

here below you'll find some comments and answers to your questions :

1. the initial conditions are the conditions you set over the entire model domain at the time t=0; the boundary conditions are the conditions you set at the boundaries of the model and that are propagated into the domain. By default, the boundary conditions are assumed constant in space and time. If you want to impose boundary conditions that are time and/or space dependent, you have to modify some subroutine of the code. I already posted a message that describes the approach to follow. You can use the same approach to impose varying boundary conditions in your model. Pay attention: this approach concerns only the BOUNDARY CONDITIONS of your model. If you have a large model domain, you must take into account the wave generation due to the wind also inside your model (see the TOMAWAG guide at chapters 7.2.6 and 8.2.4).
2. If you have a single-peak wave spectrum you need to specify only the drectional_spread_1. Generally you set this variable with reference to a wave time series or to the wave conditions in the area considered.
3. The subroutine SPEINI.F is used for setting the boundary conditions. The initial conditions (the conditions at time t=0 over the whole mesh) are set via the subroutine CONDIW.f.

Hope it helps,

Giovanni

another thing to ask about
to generate waves using unsteady wind, i will use the windx and windv data in selafin file. after reviewing the two subroutines LECDOI and NOUDON, I find they do the same thing for the wind data.
first i think i must not modify any subroutine to interpolate the wind between two time steps from the selafin file. Am I right?
what is the difference between LECDOI and NOUDON subroutines as they do the same task?


thanks

hello,
could you please give me specific answer to

what is the difference between boundary mean fetch value and initial mean fetch value? And how can I determine the fetch length for the case of constant wind, and the other one for wind that change in space and time?



realy thanks for your effort

Hello,

to answer to your last 2 post :

• LECDOI and NOUDON use the same appraoch. The difference between the two subroutines is that the LECDOI is used to initialize the simulation (t=0) and NOUDON is used during the loop over the times steps (t>0). If you use wind files in the Selaphin format, you are right, you shouldn't need to interpolate the wind fields.
• As I explained in the previous post, the initial conditions are used to set the initial spectral variance density over the whole mesh, while the boundary conditions are used to impose the sea-state conditions at the boundaries. If you set TYPE OF INITIAL DIRECTIONAL SPECTRUM to 1, you need the initial mean fetch value to define the initial sea-state conditions over the whole domain. If you set TYPE OF BOUNDARY DIRECTIONA L SPECTRUM to 1, you need to set the boundary mean fetch value to define the boundary sea-state conditions. The value you set for the fetch either is taken from wind fields maps, or is based on your assumptions. The fetch is the wind-exposed distance over the water surface, measured in the upwind direction. Either you assume this distance, or you make reference to wind fields measures.
• In case you want to set boundary conditions starting from the fetch of wind fields varying in space and time, the problems becomes more complicated. In this case, I suggest you to change the TYPE OF BOUNDARY DIRECTIONAL SPECTRUM and impose directly, as input parameters, the time series of the significant wave height and of the peak frequency taken from a larger spectral wave model (a sea-state reanalysis or a spectral wave model covering a much larger area).

Kind regards,

Giovanni

hello,
the two subroutines LECDOI and NOUDON use the variables TV1 nad TV2 for the time reference inside the serafin wind file. these two variables are the difference between the begin time of the simulation "DDC" and the wind recoded time "ATT".
- please give me a simple flowchart or idea of each step inside the TEMP.f subroutine since i can not understand how it convert dates?
- is there a time reference for the variable DDC? in order to be converted
- how the subroutine TEMP.f will work with the following example?
the begin of the simulation"DDC": 2006-01-01 00:00:00 yyyy-MM-DD HH:mm:ss
wind record time in the serafin file: 3600 i.e after one hour of the simulation begin.

thanks and kind regards

ok,
after review of the manual the DDC format is as follow YYMMDDHHmm which will be used by TEMP.f sub routine. on the other side TV1 and TV2 will be multiplied by 1.D2 to be converted to double. please give me an answer about
the following
what is the format of time to be stored inside the serafin file?

thanks

All the TOMAWAC simulations start at time t=0
The date of computation beginning is optional information included in the header of the Selaphin file, and it is used in case you use wind field files having the time in the WAM format
The subroutine TEMP calculates the difference between the DDC and a time step of the wind field file, only in the case that the time in the wind file has the format as in the WAM formulation : yymmddhhmm (TOMAWAC manual, chapter 8.1.3)
For your example:
DDC= 0601010000
The time in the wind field corresponding to 3600s should be written, therefore, as 0601010100
The TEMP subroutines gives as output TV=3600, which corresponds to one hour after the computation beginning.

If the time in your Selaphin wind field file has a different format (e.g. expressed in seconds and starting from time t=0), I suggest you to modify the subroutines NOUDON and LECDOI, in order to read directly the time TV, without using the subroutine TEMP. In this case you do not need to use your DDC. The code should be modified in the following way

If you define the data in the wind field file with a different format, you should modify the same soubroutines, in order to compare properly the TOMAWAC computation time and the wind field time.

It should work.

Regards,

Giovanni

hello,

to complete my understanding i should to modify the code as you give me in the previous post, because of the reason that may occur in the following situation:
if i want to use the time format inside the serafin file as yymmddhhmm. the time value will be saved as single precision real not double, so it will be truncated for example:
if the time is 9601010015 it will be saved inside the serafin as 96010100. while after the read by noudon.f subroutine it will be multiplied by 1.D2 and will be 9601010000. which may give error during the interpolation, since it is not the requested TV2=9601010015.

please feel free to correct my understanding,

thanks and kind regards

hello,
another question
if i use the linear wave growth, so i will need no initial wave spectrum inside my domain. so which keywords will be used with linear wave growth for the initial conditions inside my domain?
can i use together, TYPE OF BOUNDARY DIRECTIONAL SPECTRUM equal to 1 for my boundaries and the linear wave growth inside ?

thanks