Garden Sprinkler effect

large Δθ
small Δθ
Swan control file

Large directional steps

In oceanic circumstances a concentrated storm will generate a very local wind sea. The wave energy is distributed over a large range of spectral directions which subsequently will propagate away from the generation area as swell. In a numerical model the energy is not spread over a continuous range of directions but over discrete directions. Far away from the source area this leads to a number of discrete energy blobs. The area shown below is 1000km ∗ 800km. The computation extends over 2 days; the time step in the animation is 1 hr, but the computational time step was 20 min. Further details of the computation can be found in the control file. The directional step size is 10° which is quite sufficient for the wind sea as such. The quantity shown is the significant wave height.

Note that the color scaling changes during the animation. In other words a certain color does not correspond to the same wave height in different pictures. The Swan control file which generates this animation is shown below.

go to top

Small directional step

For comparison another computation is made with a resolution of 2°. Now the wave field does not disintegrate any more.

go to top

Swan control file

In the file below the initial wave field is obtained by means of a wind field; after this the wind source term is switched off. Also the other source terms are made 0. The bottom is horizontal so there is no refraction. The depth is 100m everywhere; this is achieved by the factor 100. in the command READINP BOTTOM; the file reads:
 1.  1.
 1.  1.
The initial wave field is obtained by introducing a simple wind field in the upper left corner of the computational area. The file A24.wnd reads:
0  0  0
0 14  0
0  0  0
0  0  0
0 -9  0
0  0  0
As this input illustrates it is not necessary to have a bottom and/or a wind grid with the same (spatial) resolution and boundaries as the computational grid. In practical applications this is useful because the wind field changes much slower than the bottom level so that the steps for the wind grid can be larger.
!*******************  HEADING  ************************************

Project 'SWAN Conf. tests' 'A24a'

! PURPOSE OF TEST: Garden Sprinkler effect
! A24a: 2d regular grid computation

!   +--------------------------------------------------------+
!   | This SWAN input file is on of a set of configuration   |
!   | tests for SWAN.                                        |
!   +--------------------------------------------------------+

!*****************  MODEL INPUT  ****************************

mode nonstat
coord cartesian

CGRID 0. 0. 0. 1000000. 800000. 100 80 Circle 36 0.05 0.25 20

INPGRID BOTTOM 0. 0. 0. 1 1 1000000. 1000000.

inputgrid wind regular 0. 700000. 0. 2 2 50000. 50000.
read wind 1. 'A24.wnd' idla=1 nhedvec=1
! initial wave field derived from wind (local peak)
init default

off windgrowth
off wcap
off quad

!***************** OUTPUT REQUESTS *************************

Graphics POSTscript
palette pos 1.,1.,0.4, 0.,1.,0., 0.,0.,1., 1.,0.,0., 0.3,0.,0.
PLOT  'COMPGRID' FILE  ''  color HS  &
        output 20080101.0000  2 hr
COMPUTE nonstat 20080101.0000  20 min 20080103.0000

go to top

© 2012: Nico Booij