Clarification needed on the type of structured grid required for a standalone WW3 run

[xuchengov]

Before generating the structured grid, we must define what type of structured grid is required for WW3 testing:

Options:

Option A: Regular Latitude-Longitude Grid
Simple rectangular grid with uniform spacing (fixed resolution)
Poor coastal alignment, but suitable for replicable and controlled standalone runs
What resolution should we select?

Option B: Orthogonal Curvilinear Grid
Structured grid with grid lines aligned to coastline curvature
Better representation of coastal boundary and wetting/drying zones
Requires more advanced mesh generation tools and configuration

[yunfangsun]

Due to the stoke drift:

we would like to try to run WW3 in structured grid.

old mesh

first tried

[yunfangsun]

The original mesh has 23018 nodes,

After talk with @xuchengov, we decided not to fit both the northern and southern boundaries (the magenta nodes), and with a 100m resolution the mesh will looks like:

it will have 49650 nodes,

200m resolution:

it will have 11780 nodes,

300m resolution:

it will have 5550 nodes.

Here, I plan to use 200m resolution mesh to do the structured mesh test.

[yunfangsun]

Hi @xuchengov,

Thank you very much! I got your structured ww3 configuration, it is strictly along longitude and latitude with the ww3_gs_xcoord.dat and ww3_gs_ycoord.dat defined the nodes's x and y coordinates, it looks like

And this configuration don't fit our scenario.

Then I dig into the manual of ww3

There are 3 type of grid in WW3:

Feature	Structured (Type 1)	Curvilinear (Type 2)	Unstructured (Type 3)
Grid alignment	Aligned with latitude and longitude	Curved or rotated grid (We planned)	Any orientation or shape
Rotation support	Not supported	Intended to support rotation	Fully supports rotation
Variable spacing	Not supported	Intended to support it	Fully supported
Supported solver	Fully implemented	N/A	Fully implemented
Numerical scheme	Finite difference	N/A	Finite volume

Your configuration is type 1, our old duck case is type 3, and it seems that type 2 could be run with the unstructured solver

[yunfangsun]

Hi @aliabdolali ,

I plan to use the following grid to test the structured ww3 code for the duck region , our goal is to compare the stroke drift velocity's behavior in structured ww3 and unstructured ww3.

And from the manual, it seems that the ww3 structured solver can not deal with the rotated grid (type 2) other than the aligned with latitude and longitude (type 1), if i want to run with the above grid, I have run it with the unstructured ww3 solver, is my understanding correct or not?

and do you have any rotated grid test case or regression test for ww3?

Thank you very much!

[yunfangsun]

From WW3 manual:

Regarding the user interface: in ww3 grid.inp, a string is used to indicate the grid type.
In cases where this grid string is ‘RECT’, the model processes input for a regular grid.
In case where this grid string is ‘CURV’, the model processes input for an irregular grid.
This Triangular unstructured grids option is activated by setting the grid string to ‘UNST’.

[aliabdolali]

can you try the curvilinear option?

[yunfangsun]

WW3 stoke drift velocity

[yunfangsun]

can you try the curvilinear option?
Hi @aliabdolali ,

Since our goal is to try resolve the issues about the stoke drift velocity differences between WW3 and WWM (schism's wind wave model), and in the big wave tag-up meeting, people believe that the structured ww3 could be a baseline. Therefore, we would like to run a structured ww3 case for the duck region.

And based on the topography and wave characteristics of the duck region, and have to use the curvilinear option. do you think the solver of CURV is solid as the RECT? and the CURV could be the baseline for this duck case?

Thank you very much!

[aliabdolali]

I think so, the other option is to rotate it and do it in xy cartesian, but I assume meshcap cannot support cartesian, please check with Ufuk.

[yunfangsun]

I think so, the other option is to rotate it and do it in xy cartesian, but I assume meshcap cannot support cartesian, please check with Ufuk.

Hi @aliabdolali ,

Thank you!, i am firstly testing the duck case with ww3 standalone outside of ufs-coastal. @uturuncoglu updated that the cartesian grid is using another meshcap other than the current one.

[yunfangsun]

Now, I am using a 150m*150m rotated grid to test the duck case, which has 20298 nodes which is very similar to the unstructured grid (23018) , the new grid is as follows:

And interpolated the bathymetry from the unstructured grid to the structured one:

[yunfangsun]

The new grid is a 199*102 node system:
And the ww3_grid successfully generated the mod_def.ww3, the log file is:

                   *** WAVEWATCH III Grid preprocessor ***
               ===============================================

  Grid name : FRF


  Spectral discretization :
 --------------------------------------------------
       Number of directions        :  36
       Directional increment (deg.):  10.0
       First direction       (deg.):   0.0
       Number of frequencies       :  35
       Frequency range        (Hz) :   0.0400-1.0219
       Increment factor            :   1.100


  Model definition :
 --------------------------------------------------
       Dry run (no calculations)   :  ---/NO
       Propagation in X-direction  :  YES/--
       Propagation in Y-direction  :  YES/--
       Refraction                  :  YES/--
       Current-induced k-shift     :  YES/--
       Source term calc. and int.  :  YES/--


  Time steps :
 --------------------------------------------------
       Maximum global time step      (s) :   60.00
       Maximum CFL time step X-Y     (s) :   60.00
       Maximum CFL time step k-theta (s) :   60.00
       Minimum source term time step (s) :   60.00

  Preprocessing namelists ...
  Preprocessing namelists finished.


  Seeding as proxi for linear input.


  Wind input (WAM 4+) (default values) :
 --------------------------------------------------
       minimum Charnock coeff.     :    0.0095
       betamax                     :    1.430
       power of cos. in wind input :    2.000
       z0max                       :    0.000
       zalp                        :    0.006

....

       Number of latitudes       :       199
       Number of grid points     :     20298
       Number of sea points      :     19899 (98.0%)
       Number of input b. points :       199
       Number of land points     :       199
       Number of excluded points :       200


  End of program
 ========================================
         WAVEWATCH III Grid preprocessor

 SI nbPlus=       81192  nbMinus=           0  nbZero=           0

  Summary grid statistics :
 --------------------------------------------------
       Number of longitudes      :       102
       Number of latitudes       :       199
       Number of grid points     :     20298
       Number of sea points      :     19899 (98.0%)
       Number of input b. points :       199
       Number of land points     :       199
       Number of excluded points :       200

End of program

     WAVEWATCH III Grid preprocessor

SI nbPlus= 81192 nbMinus= 0 nbZero= 0

Summary grid statistics :

   Number of longitudes      :       102
   Number of latitudes       :       199
   Number of grid points     :     20298
   Number of sea points      :     19899 (98.0%)
   Number of input b. points :       199
   Number of land points     :       199
   Number of excluded points :       200

End of program

     WAVEWATCH III Grid preprocessor

SI nbPlus= 81192 nbMinus= 0 nbZero= 0
SI nbPlus= 81192 nbMinus= 0 nbZero= 0

[yunfangsun]

The boundary condition is defined as follows:

! -------------------------------------------------------------------- !
! WAVEWATCH III - ww3_bounc.nml - Boundary input post - processing     !
! -------------------------------------------------------------------- !


! -------------------------------------------------------------------- !
! Define the input boundaries to preprocess via BOUND_NML namelist
!
! *namelist must be terminated with /
! *definitions & defaults:
!     BOUND%MODE = 'WRITE'            ! ['WRITE' | 'READ']
!     BOUND%INTERP = '2'                  ! interpolation[1(nearest), 2(linear)]
!     BOUND%VERBOSE = '1'                  ! [0 | 1 | 2]
!     BOUND%FILE = 'spec.list'        ! input_spec.nc listing file
! -------------------------------------------------------------------- !
&BOUND_NML
BOUND%MODE = 'WRITE'                             ! ['WRITE' | 'READ'] !
BOUND%INTERP = 2                ! interpolation[1(nearest), 2(linear)] !
BOUND%VERBOSE = 1                                        ! [0 | 1 | 2] !
BOUND%FILE = spec.list                    ! input_spec.nc listing file !
/
! -------------------------------------------------------------------- !
! WAVEWATCH III - end of namelist                                      !
! -------------------------------------------------------------------- !

It also successfully got the nest.ww3

[yunfangsun]

the restart file definition:
$ -------------------------------------------------------------------- $
$ WAVEWATCH III Initial conditions input file $
$--------------------------------------------------------------------- $
$ type of initial field ITYPE .
$
5
$
$ ITYPE = 1 ---------------------------------------------------------- $
$ Gaussian in frequency and space, cos type in direction.
$ - fp and spread (Hz), mean direction (degr., oceanographic
$ convention) and cosine power, Xm and spread (degr. or m) Ym and
$ spread (degr. or m), Hmax (m) (Example for lon-lat grid in degr.).
$
$ 0.10 0.01 270. 2 1. 0.5 1. 0.5 2.5
0.10 0.01 270. 2 0. 1000. 1. 1000. 2.5
$ 0.10 0.01 270. 2 0. 1000. 1. 1000. 0.01
$ 0.10 0.01 270. 2 0. 1000. 1. 1000. 0.
$
$ ITYPE = 2 ---------------------------------------------------------- $
$ JONSWAP spectrum with Hasselmann et al. (1980) direct. distribution.
$ - alfa, peak freq. (Hz), mean direction (degr., oceanographical
$ convention), gamma, sigA, sigB, Xm and spread (degr. or m) Ym and
$ spread (degr. or m) (Example for lon-lat grid in degr.).
$ alfa, sigA, sigB give default values if less than or equal to 0.
$
$ 0.0081 0.1 270. 1.0 0. 0. 1. 100. 1. 100.
$
$ ITYPE = 3 ---------------------------------------------------------- $
$ Fetch-limited JONSWAP
$ - No additional data, the local spectrum is calculated using the
$ local wind speed and direction, using the spatial grid size as
$ fetch, and assuring that the spectrum is within the discrete
$ frequency range.
$
$ ITYPE = 4 ---------------------------------------------------------- $
$ User-defined spectrum
$ - Scale factor., defaults to 1 if less than or equal 0.
$ - Spectrum F(f,theta) (single read statement)
$
$ -0.1
$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
$ 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
$ 0 1 4 2 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
$ 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
$ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
"ww3_strt.inp" 72L, 3398B

It gave the restart.ww3