new feature: add multi faults tsunamis

Feature introduced in paper from TUC 2022. It required:

1. A txt file attached here that contains:
   • in the header the number of subfaults.
   • for each row the okada information for each subfault + dynamic information. Details below in the okada modified file:
   • was introduced as FORMATTED DATA FILE 1 = 'earthquakes/tohoku_fuji_50km_dyn10s.txt' see full CAS file 03_tokohu_2011_fuji_10s.cas
2. A modified version of okada.f with 12 coefs instead of the existing 10, in the same order:
   • TI = COEFS(1) ! TIME INIT - **NEW** DYNAMIC // TSR
   • TR = COEFS(2) ! RISING TIME - **NEW** DYNAMIC // TSR
   • HH = COEFS(3) ! FOCAL DEPTH
   • L = COEFS(4) ! FAULT LENGTH
   • W = COEFS(5) ! FAULT WIDTH
   • D = COEFS(6) ! DISLOCATION
   • TH = COEFS(7) * DEG2RAD ! STRIKE DIRECTION
   • DL = COEFS(8) * DEG2RAD ! DIP ANGLE
   • RD = COEFS(9) * DEG2RAD ! SLIP ANGLE
   • Y0 = COEFS(10) * DEG2RAD ! EPICENTRE LATITUDE
   • X0 = COEFS(11) * DEG2RAD ! EPICENTRE LONGITUDE
   • C0 = COEFS(12) ! SIZE OF THE ELLIPSE
3. A user input file to read the multi-faults info:
   • here is the dynamic implementation, i.e. modifying bottom over time: user_utimp_telemac2d.f
   • here is the hydrostatic implementation, i.e. translating okada fault deformation into free-surface anomaly: user_condin_h.f.
   • the best way forward would be to implement static when the first 2 parameters are equal to 0 (i.e. TIME INIT = 0 and RISING TIME = 0)

It is the first time I implement a new feature, that requires also new keywords: (we should add a new keyword like INPUT FILE FOR TSUNAMI CHARACTERISTICS or TSUNAMI INPUT FILE?)

@chi-tuan.pham, please have a look at the files and let me know what you think + how to proceed:

1. to add new keyword (+ suppress the old one?)
2. static vs dynamic implementation (mentioned above)
3. add a new test case (or modify the existing okada test case)

also pinging @sebastien.bourban, @michael.turnbull and @alexander.breugem for more thoughts on this

added lang:fortran module:telemac-2d module:telemac-3d type:feature labels

assigned to @saillour.thomas

changed the description

changed the description

Hello Thomas,

I am hesitating between 3 names:

• TSUNAMIS INPUT FILE,
• TSUNAMIS FILE,
• TSUNAMIS DATA FILE. I let you other participants give their opinion. Perhaps the shorter, the better if no other keywords deal with tsunamis.

I wonder if the old keyword should be deleted or not, it is quite simple to use if only 1 single fault tsunami and no revolution of use for current users (except adding 2 extra coefficients). The discrimination could be done if the "TSUNAMIS FILE" or another name is filled with a non void file name: in that case, read the file. Otherwise use the old way. To be derived in CONDIN subroutine for the computation of initial conditions.

For the order of coefficients, I wonder if the 2 new variables TI and TR should not be added at the end and not before the historic ones (from 11 to 12 rather than 1 to 2). Once more, I let you and other users give their opinion.

In the implementation of CONDI_OKADA of the OKADA module, I think you can let the old name for variable H rather than H letting your comment in the headline + perhaps a comment when first used in the code. Do not forget to add the description of AT and LT in the header of CONDI_OKADA subroutine.

USER_CONDIN_H to be merged with CONDIN subroutine. I am not sure writing every information you have written is necessary, in particular if NLINES is big. Do not forget to deallocate COEFS if allocated. Be careful, the dimension of 2nd component of COEFS is 10, not 12.

For your USER_UTIMP_TELEMAC2D part, it should be implemented directly in TELEMAC2D subroutine, but:

• the reading of COEFS should be done only once, if not already done for the initial conditions ;
• COEFS should be deallocated in DEALL_TELEMAC2D,
• quite strange to let a suroutine which name starts from CONDI_ is called at every time step in case of dynamic version, perhaps rename the subroutine, I do not have an obvious idea. Perhaps you or the other participants have better ideas?
• to prevent from too many writings, delete or comment lines

WRITE(LU,*) & 'MODIFYING BOTTOM WITH DYNAMIC OKADA - SUBFAULT',LINE

For the example, you can build another one based on the existing okada or build another one with another geometry if CPU time is not too long.