** Warnin: Quadruplets will be updated per iteration

[manhnt181]

Hi everyone,
I'm a new user of the COAWST model. I've been using WRF+ROMS+SWAN option, however, I got the warning when running the model:

** Warning : Quadruplets will be updated per iteration
** Warning : instead of per sweep. This will increase
** Warning : amount of internal memory with a factor 2

I don't exactly this warning maybe cause any problem, but I got another error in the log.file:

Fatal error in MPI_Recv: Other MPI error, error stack:
MPI_Recv(186).............: MPI_Recv(buf=0x7ffd3d7a59bc, count=1, MPI_INTEGER, src=0, tag=0, comm=0x84000002, status=0x47a6940) failed
dequeue_and_set_error(596): Communication error with rank 0

Here is my SWAN's namelist:

Code: Select all

&& KEYWORD for number of nested SWAN grids.
NSGRIDS 1

&& KEYWORDS TO CREATE AND READ COMPUTATIONAL GRID &&
CGRID CURVILINEAR 389 284 EXC 9.999000e+003 9.999000e+003 CIRCLE 36 0.04 1.0 24
READGRID COORDINATES 1 'Projects/Haima/swan_coord.grd' 4 0 0 FREE

&& KEYWORDS TO CREATE AND READ BATHYMETRY GRID &&
INPGRID BOTTOM CURVILINEAR 0 0 389 284 EXC 9.999000e+003
READINP BOTTOM 1 'Projects/Haima/swan_bathy.bot' 4 0 FREE

&& KEYWORD TO CREATE CURRENT GRID &&
INPGRID CURRENT CURVILINEAR 0 0 389 284 EXC 9.999000e+003 &
       NONSTATIONARY 20161017.000000 6 DAY 20161023.000000

&& KEYWORD TO CREATE WATER LEVEL GRID &&
INPGRID WLEV CURVILINEAR 0 0 389 284 EXC 9.999000e+003 &
       NONSTATIONARY 20161017.000000 6 DAY 20161023.000000

&& KEYWORD TO CREATE BOTTOM FRIC GRID &&
INPGRID FRIC CURVILINEAR 0 0 389 284 EXC 9.999000e+003 &
       NONSTATIONARY 20161017.000000 6 DAY 20161023.000000

&& KEYWORD TO CREATE WIND GRID &&
INPGRID WIND REGULAR 96.9742 -2.3665 0 389 285 0.2 0.2 EXC 9.999000e+003 &
      NONSTATIONARY 20161017.000000 6 HR  20161023.000000
&READINP WIND 1 'Projects/Sandy/swan_GFS_Oct2012.dat' 4 0 FREE
& Boundary files  ****************************************
& 2D Spec Boundary files  ****************************
BOUND SHAPESPEC JONSWAP MEAN DSPR DEGREES
BOUNDSPEC SEGMENT IJ    0    0    0   284 CONSTANT PAR  0.1  20.0    0.0  15
BOUNDSPEC SEGMENT IJ    0    0  389     0 CONSTANT PAR  0.1  20.0    0.0  15
BOUNDSPEC SEGMENT IJ  389    0  389   284 CONSTANT PAR  0.1  20.0    0.0  15
BOUNDSPEC SEGMENT IJ    0  284  389   284 CONSTANT PAR  0.1  20.0    0.0  15
& **   COPY THESE LINES TO SWAN INPUT FILE   ********
& 2D Spec Boundary files  ****************************
& Restart name **********************************
&INIT
INIT_DEFAULT
& PHYSICS  ****************************************
BREAKING CONSTANT 1.0 0.73
FRICTION MADSEN 0.05
&OFF QUAD
GEN3 KOMEN
WCAP KOM 2.36E-5 3.02E-3 2.0 1.0 1.0
PROP BSBT
BLOCK 'COMPGRID' NOHEADER 'Haima_hsig.mat'   LAY 4 HSIGN 1.  OUTPUT 20161017.000000 1 HR
BLOCK 'COMPGRID' NOHEADER 'Haima_xp.mat'     LAY 4 XP 1.
BLOCK 'COMPGRID' NOHEADER 'Haima_yp.mat'     LAY 4 YP 1.
BLOCK 'COMPGRID' NOHEADER 'Haima_wind.mat'   LAY 4 WIND 1.   OUTPUT 20161017.000000 1 HR
BLOCK 'COMPGRID' NOHEADER 'Haima_hsig.nc'    LAY 4 HSIGN 1.  OUTPUT 20161017.000000 1 HR
RESTART 'swan_rst.dat' FREE 1 HR

&COMPUTE STAT 20161017.000000
COMPUTE NONSTAT 20161017.000000 1 HR 20161023.000000

HOTFILE 'Haima_init.hot'

STOP

Please help me to solve this problem.
Any suggestions are welcome and appreciated!

[jcwarner]

for the quadruplets, swanmain.F says

"!
! Here the various problems with quadruplets are checked
! The combination of quadruplets and sectors is an error
! in the calculation of quadruplets when the SECTOR option is
! used in the CGRID command. This error should be corrected
! in the future
!
IF (ICUR.NE.0 .AND. (IQUAD.EQ.1 .OR. IQUAD.EQ.2)) THEN 41.55
CALL MSGERR(1,'Quadruplets will be updated per iteration') 41.55
CALL MSGERR(1,'instead of per sweep. This will increase') 41.55
CALL MSGERR(1,'amount of internal memory with a factor 2') 41.55
IQUAD = 3 41.55
ENDIF 41.55
!"

you can deactivate using
OFF QUAD

For the MPI error, that is not enough information there. the error probably happened earlier.

[manhnt181]

Hi John,
Thank you for your suggestion!
I've changed my SWAN's namelist with "OFF QUAD" option, but I got this error:

&COMPUTE STAT 20161017.000000
COMPUTE NONSTAT 20161017.000000 1 HR 20161023.000000
** Error : Quadruplets should be activated when SWAN
** Error : is running in a third generation mode and
** Error : wind is present
** Warning : Limiter is de-activated in case of no quadruplets
** No start of computation because of error level: 2
** To ignore this error, change [maxerr] with the SET command

[manhnt181]

Hi John,
In term of the warning (Quadruplets will be updated per iteration), I've fixed it by changed my grid points in SWAN's namelist and my script.in. To be more detail, I've increase the number of grid point of SWAN model by 1 to equal with the number of grid point of ROMS model.
- My ROMS's domain has 390x285 grid points:

>> netcdf_load('ROMS_COAWST_grd1.nc')
>> size(h)
ans =
390 285

- So I change my SWAN's namelist like this:

PROJECT 'Haima' '1'
'3Way coupling WRF-SWAN-ROMS'
''
''
MODE NONSTATIONARY TWODIMENSIONAL
SET DEPMIN 0.10 INRHOG 1 NAUTICAL PWTAIL 5
COORDINATES SPHERICAL

&& KEYWORD for number of nested SWAN grids.
NSGRIDS 1

&& KEYWORDS TO CREATE AND READ COMPUTATIONAL GRID &&
CGRID CURVILINEAR 390 285 EXC 9.999000e+003 9.999000e+003 CIRCLE 36 0.04 1.0 24
READGRID COORDINATES 1 'Projects/Haima/swan_coord.grd' 4 0 0 FREE

&& KEYWORDS TO CREATE AND READ BATHYMETRY GRID &&
INPGRID BOTTOM CURVILINEAR 0 0 390 285 EXC 9.999000e+003
READINP BOTTOM 1 'Projects/Haima/swan_bathy.bot' 4 0 FREE

&& KEYWORD TO CREATE CURRENT GRID &&
INPGRID CURRENT CURVILINEAR 0 0 390 285 EXC 9.999000e+003 &
NONSTATIONARY 20161017.000000 6 DAY 20161022.000000

&& KEYWORD TO CREATE WATER LEVEL GRID &&
INPGRID WLEV CURVILINEAR 0 0 390 285 EXC 9.999000e+003 &
NONSTATIONARY 20161017.000000 6 DAY 20161023.000000

&& KEYWORD TO CREATE BOTTOM FRIC GRID &&
INPGRID FRIC CURVILINEAR 0 0 390 285 EXC 9.999000e+003 &
NONSTATIONARY 20161017.000000 6 DAY 20161023.000000
& Boundary files ****************************************
& 2D Spec Boundary files ****************************
BOUND SHAPESPEC JONSWAP MEAN DSPR DEGREES
BOUNDSPEC SEGMENT IJ 0 0 0 285 CONSTANT PAR 0.1 20.0 0.0 15
BOUNDSPEC SEGMENT IJ 0 0 390 0 CONSTANT PAR 0.1 20.0 0.0 15
BOUNDSPEC SEGMENT IJ 390 0 390 285 CONSTANT PAR 0.1 20.0 0.0 15
BOUNDSPEC SEGMENT IJ 0 285 390 285 CONSTANT PAR 0.1 20.0 0.0 15
& Restart name **********************************

&INIT
INIT_DEFAULT

& PHYSICS ****************************************
BREAKING CONSTANT 1.0 0.73
FRICTION MADSEN 0.05
&OFF QUAD
GEN3 KOMEN
WCAP KOM 2.36E-5 3.02E-3 2.0 1.0 1.0
PROP BSBT

BLOCK 'COMPGRID' NOHEADER 'Haima_hsig.mat' LAY 4 HSIGN 1. OUTPUT 20161017.000000 1 HR
BLOCK 'COMPGRID' NOHEADER 'Haima_xp.mat' LAY 4 XP 1.
BLOCK 'COMPGRID' NOHEADER 'Haima_yp.mat' LAY 4 YP 1.
BLOCK 'COMPGRID' NOHEADER 'Haima_wind.mat' LAY 4 WIND 1. OUTPUT 20161017.000000 1 HR
BLOCK 'COMPGRID' NOHEADER 'Haima_hsig.nc' LAY 4 HSIGN 1. OUTPUT 20161017.000000 1 HR
RESTART 'swan_rst.dat' FREE 1 HR

&COMPUTE STAT 20161017.000000
COMPUTE NONSTAT 20161017.000000 180 SEC 20161023.000000

HOTFILE 'Haima_init.hot'

STOP

and my script.in is:

SWAN_COORD(1)='../../Projects/Haima/swan_coord.grd',
SWAN_BATH(1)='../../Projects/Haima/swan_bathy.bot',
SWAN_NUMX(1)=390,
SWAN_NUMY(1)=285,
CARTESIAN(1)=0,

The size of ROMS model is:

Lm == 388 ! Number of I-direction INTERIOR RHO-points -2
Mm == 283 ! Number of J-direction INTERIOR RHO-points -2
N == 16 ! Number of vertical levels
ND == 0 ! Number of wave directional bins

However, I got the error:

MCT::m_SparseMatrixPlus::initFromRoot_:: FATAL--length of vector y different from row count of sMat.Length of y = 111826 Number of rows in sMat = 111150

Please help me to solve this problem.
Thank you very much!

[jcwarner]

this is becoming non-roms related, so can you post this message here:
https://github.com/jcwarner-usgs/COAWST/issues
thanks,
j

[manhnt181]

Hi John,
Thank you for your reply. I've posted my issue on your site.
Please help me to solve it.
Thanks,