Collab with Ali Abdolali on WW3 + SCHISM coupling & validation

[janahaddad]

Meeting notes from Tues May 7 2024

WW3 Validation

Field output

• Hs, Tp (as Fp), Mean period T01, mean Theta, Mean Wind Speed
  • Hs and wind speed are the typ satellite altimetry obs
  • Others will be useful for analysis
• ounf inp file
  • set Line 32 flag to F. This saves separate nc files for each variable and allows us to save space in Ali's GDrive
  • see /inp directory in WW3 for an example file, just need to change L32 to F

Point output

• use the example ounp int file in WW3 repo as-is (changing only top lines for your simulation start, etc.)
  • Line 33: Type 1 & subtype 3
• want full spectra for these point outputs and then derive the stats & do our own swell/wind split, etc, then compare with gauge obs

Global mesh

• Ali has a mesh from which BCs can be extracted and will share
• Yunfang will share coordinates of coupling mesh open boundary nodes
• Forcing for the global run: ERA5 winds

ERA5 to generate BCs for the Ian case:

• will use available wave stats in the wave renalaysis dataset to estimate/reconstruct spectra from parametric spectra e.g. JONSWAP or P-M
  • Ali's wave tools also do this
  • would need to get both swell & wind sea data, reconstruct separately and combine spectra

mtg outcomes:

Three possible avenues for comparison:

1. reconstruct BCs from ERA5 "reanalyzed" waves
2. extract spectral BCs from the full global run: Run global UnWW3 to provide boundary to Atlantic set up WW3#9
3. run w/o OBCs

Want to prioritize the coupling aspect, so hold off on 1 for now.
Focus on global run in the next week.

• global run needs 2 week spin-up to properly propagate swell into Atl domain per Ali prior studies

For run w/o BCs, Yunfang needs to re-run to generate spectra at poitns, and save only variables above in field output

cc @yunfangsun @aliabdolali @saeed-moghimi-noaa

[janahaddad]

@aliabdolali can you confirm T01 is ok for mean period, or would you like to have the other mean T estimates (T02 and/or T0M1)

[aliabdolali]

depends where is our emphasize, if we want to look at mean period of higher frequencies, I'd go with T02, but it is good as of now. we can calculate all of them from full spectrum at bouys, while satellite data does not contain period info, so it is not used there. It is for visualization purpose. If you plot Tp, you would see the field plots jumpy, therefore we would go with mean period.

[aliabdolali]

Meeting notes from Tues May 7 2024

WW3 Validation

Field output

• Hs, Tp (as Fp), Mean period T01, mean Theta, Mean Wind Speed

  • Hs and wind speed are the typ satellite altimetry obs
  • Others will be useful for analysis

• ounf inp file

  • set Line 32 flag to F. This saves separate nc files for each variable and allows us to save space in Ali's GDrive
  • see /inp directory in WW3 for an example file, just need to change L32 to F

Point output

• use the example ounp int file in WW3 repo as-is (changing only top lines for your simulation start, etc.)

  • Line 33: Type 1 & subtype 3

• want full spectra for these point outputs and then derive the stats & do our own swell/wind split, etc, then compare with gauge obs

Global mesh

• Ali has a mesh from which BCs can be extracted and will share
• Yunfang will share coordinates of coupling mesh open boundary nodes
• Forcing for the global run: ERA5 winds

ERA5 to generate BCs for the Ian case:

• will use available wave stats in the wave renalaysis dataset to estimate/reconstruct spectra from parametric spectra e.g. JONSWAP or P-M

  • Ali's wave tools also do this
  • would need to get both swell & wind sea data, reconstruct separately and combine spectra

mtg outcomes:

Three possible avenues for comparison:

1. reconstruct BCs from ERA5 "reanalyzed" waves
2. extract spectral BCs from the full global run
3. run w/o OBCs

Want to prioritize the coupling aspect, so hold off on 1 for now. Focus on global run in the next week.

• global run needs 2 week spin-up to properly propagate swell into Atl domain per Ali prior studies

For run w/o BCs, Yunfang needs to re-run to generate spectra at poitns, and save only variables above in field output

cc @yunfangsun @aliabdolali @saeed-moghimi-noaa

nice summary, we need more of you Jana, well organized.

[janahaddad]

thx @aliabdolali

[aliabdolali]

@janahaddad @yunfangsun @saeed-moghimi-noaa
NOS team, I uploaded the global config on our shared drive, and added the OB nodes in ww3_shel.nml as well.
Also, please change the spectral resolution consistent with your Atlantic simulations in ww3_grid.nml
You need to download ERA5 wind, and if no ICE and current are used, turn their flags to F in ww3_shel.nml and just run with wind.
then execute ww3_ounp to generate spec.
Let me know if you had questions.
Your simulation should start 15 days before Atlantic simulation and ends at the end of Atlantic simulations.

Meantime, I am working on validation.

[yunfangsun]

Hi @aliabdolali

Thank you very much, I have downloaded the configuration and ERA5 wind.

When I am running the preprocessing by ww3_grid ww3_grid.nml > ww3_grid.out, it give me the following error message in ww3_grid.out:

  Wind input (WAM 4+) (user def. values) :
 --------------------------------------------------
       minimum Charnock coeff.     :    0.0095
       betamax                     :    1.330
       power of cos. in wind input :    2.000
       z0max                       :    0.000
       zalp                        :    0.006
       Height of input wind (m)    :   10.00
       wind stress sheltering      :    1.000
       swell attenuation param.    :    1
       swell attenuation factor    :    0.800
       swell attenuation factor2   :   -0.018
       swell attenuation factor3   :    0.015
       critical Reynolds number    : 100000.0
       swell attenuation factor5   :    1.200
       swell attenuation factor6   :    0.000
       swell attenuation factor7   :    230000.000
       ratio of z0 for orb. & mean :    0.040


  Nonlinear interactions (DIA) (user def. values) :
 --------------------------------------------------
       Lambda                      :    0.25
       Prop. constant              : 0.250E+08
       kd conversion factor        :    0.75
       minimum kd                  :    0.50
       shallow water constants     :    5.50  0.83 -1.25


 *** WAVEWATCH III ERROR IN READNL :
     ERROR IN READING SDS4  IOSTAT =      18

And I see it in namelists.nml

SNLCS1 = 5.500, SNLCS2 = 0.833, SNLCS3 = -1.250 /
&SDS4 SDSBCHOICE = 1, SDSC2 = -0.2200E-04, SDSCUM = -0.4034E+00,
SDSC4 = 0.1000E+01, SDSC5 = 0.0000E+00, SDSC6 = 0.3000E+00,
WNMEANP =0.50, FXPM3 =4.00,FXFM3 =9.90,
SDSBINT = 0.3000E+00, SDSBCK = 0.0000E+00, SDSABK = 1.500, SDSPBK = 4.000,
SDSHCK = 1.50, SDSBR = 0.9000E-03, SDSSTRAIN = 0.000,
SDSP = 2.00, SDSISO = 2, SDSCOS =2.0, SDSDTH = 80.0,
SDSBRF1 = 0.50, SDSBRFDF = 0,
SDSBM0 = 1.00, SDSBM1 = 0.00, SDSBM2 = 0.00, SDSBM3 = 0.00, SDSBM4 = 0.00,
, WHITECAPWIDTH = 0.30/

Could you please give me suggestions to fix this error?

Thank you very much!

[aliabdolali]

what version of WW3 are you using? what switch?
could you compare your switch with
PDLIB SCOTCH NOGRB DIST MPI PR3 UQ FLX0 SEED ST4 STAB0 NL1 BT1 DB1 MLIM FLD1 TR0 BS0 WNX1 WNT1 CRX1 CRT1 O0 O1 O2 O3 O4 O5 O6 O7 O14 O15 IC0 IS0 REF0

and send me your Atlantic ww3_grid.inp(nml)

[yunfangsun]

Hi @aliabdolali
I am using the following switch for the version 8b5e91f0887ab2b8c218228fff6cb82297a613b3:
NCO PDLIB SCOTCH NOGRB DIST MPI PR3 UQ FLX0 SEED ST4 STAB0 NL1 BT1 DB1 MLIM FLD2 TR0 BS0 RWND WNX1 WNT1 CRX1 CRT1 O0 O1 O2 O3 O4 O5 O6 O7 O14 O15 IC0 IS0 REF0 SCRIP SCRIPNC

I will try the switch you suggested.

And I have just uploaded the Atlantic ww3_gird.inp to the google drive https://drive.google.com/drive/folders/1CVmyF70egGu7Ke-43bLLMYaEpGFFHl1-

Thank you!

[yunfangsun]

Hi @aliabdolali

I am using WW3 version 14e6bedfcefeff8f0896acea5dade4100a3fbacd and switch
PDLIB SCOTCH NOGRB DIST MPI PR3 UQ FLX0 SEED ST4 STAB0 NL1 BT1 DB1 MLIM FLD1 TR0 BS0 WNX1 WNT1 CRX1 CRT1 O0 O1 O2 O3 O4 O5 O6 O7 O14 O15 IC0 IS0 REF0

And the error message in ww3_grid.out is still

Nonlinear interactions (DIA) (user def. values) :
--------------------------------------------------
      Lambda                      :    0.25
      Prop. constant              : 0.250E+08
      kd conversion factor        :    0.75
      minimum kd                  :    0.50
      shallow water constants     :    5.50  0.83 -1.25


*** WAVEWATCH III ERROR IN READNL :
    ERROR IN READING SDS4  IOSTAT =      18

And I set this configuration on Hera, could you give me any further instructions to fix it?

Thank you very much!

[aliabdolali]

I just updated namelists.nml on GD
there was two ",," instead of "," and it caused an issue.

[aliabdolali]

@yunfangsun could you upload the *wnd.nc file to google drive?

[janahaddad]

just moved @yunfangsun 's download_inv.nc file. Let me know if there's an issue with it

[yunfangsun]

Thank you! @janahaddad

[yunfangsun]

Hi @aliabdolali ,

I have tried to run the buoy_sat_analysis.m, and it requires the function of get_taylor_diagram_options, and I downloaded it from https://viewer.mathworks.com/?viewer=plain_code&url=https%3A%2F%2Fit.mathworks.com%2Fmatlabcentral%2Fmlc-downloads%2Fdownloads%2Ff371bd0e-5aed-4e75-9f8c-523ae637da41%2F394a1f79-6954-41d1-b536-5034c2a18722%2Ffiles%2Fget_taylor_diagram_options.m&embed=web

And then the script requires function of check_on_off, which is called by get_taylor_diagram_options, I can't find this function on line, could you please share the taylor_diagram related matlab functions?

Thank you very much!

[aliabdolali]

have you tried this: https://github.com/PeterRochford/SkillMetricsToolbox/tree/master
All of required libraries are there.

[yunfangsun]

I have removed the plotting part, and generated the 4 csv files, and compared with @aliabdolali 's results, and they matched each other.

[yunfangsun]

https://github.com/noaa-ocs-modeling/SurgeTeamCoordination/discussions/501 --> How to run matlab interactively with GUI on Hera

[yunfangsun]

Hi @aliabdolali

I have uploaded the spectra (325 files) from the outputs of the global wave model in the folder https://drive.google.com/drive/folders/1z_944rq4k8_F5M3dSFHJIs98rf_TA6kr

Could you please take a look, thank you very much

[yunfangsun]

Hi @aliabdolali ,

I have uploaded the subset mesh to Google Drive, https://drive.google.com/drive/folders/1gFXWJn4KVlZsRIZmPa73oWIWfkOKFWy5

Could you please take a look at the mesh quality?

Thank you very much!

[aliabdolali]

could you add three meshes you used?

[yunfangsun]

Hi @aliabdolali

I have uploaded the three meshes to https://drive.google.com/drive/folders/1gFXWJn4KVlZsRIZmPa73oWIWfkOKFWy5

atlantic_adcirc.msh 1813443 nodes
atlantic_coarse.msh 557481 nodes
atlantic_subset.msh. 2470094 nodes

The mesh that has issues to start is atlantic_subset.msh

[yunfangsun]

Hi @aliabdolali ,

I have tested your suggestion by using the following ww3_shel.nml:

! -------------------------------------------------------------------- !
! WAVEWATCH III ww3_shel.nml - single-grid model                       !
! -------------------------------------------------------------------- !


! -------------------------------------------------------------------- !
! Define top-level model parameters via DOMAIN_NML namelist
! -------------------------------------------------------------------- !
&DOMAIN_NML
  DOMAIN%IOSTYP  = 1
  DOMAIN%START   = '20210801 000000'
  DOMAIN%STOP    = '20211002 000000'
/

! -------------------------------------------------------------------- !
! Define each forcing via the INPUT_NML namelist
! -------------------------------------------------------------------- !
&INPUT_NML
  INPUT%FORCING%ICE_CONC      = 'F'
  INPUT%FORCING%CURRENTS      = 'F'
  INPUT%FORCING%WINDS         = 'T'
/

! -------------------------------------------------------------------- !
! Define the output types point parameters via OUTPUT_TYPE_NML namelist
! -------------------------------------------------------------------- !
&OUTPUT_TYPE_NML
  TYPE%FIELD%LIST          = 'WND HS FP DP PTP PDIR DIR T01 SPR T02 TM10 LM'
  TYPE%POINT%FILE          = 'ww3_points.list'
/

! -------------------------------------------------------------------- !
! Define output dates via OUTPUT_DATE_NML namelist
! -------------------------------------------------------------------- !
&OUTPUT_DATE_NML
  DATE%FIELD          = '20210801 000000' '3600' '20211002 000000'
  DATE%POINT          = '20210801 000000' '3600' '20211002 000000'
  DATE%RESTART        = '20210801 000000' '86400' '20211002 000000'
/

! -------------------------------------------------------------------- !
! Define homogeneous input via HOMOG_COUNT_NML and HOMOG_INPUT_NML namelist
! -------------------------------------------------------------------- !
&HOMOG_COUNT_NML
/

&HOMOG_INPUT_NML
/
! -------------------------------------------------------------------- !
! WAVEWATCH III - end of namelist                                      !
! -------------------------------------------------------------------- !

And the log.ww3 shows:

 WAVEWATCH III log file                                 version 7.14
 =====================================================================
                                                  date : 2024/12/03
                                                  time :  09:47:06

 Model definition file read.
 Restart file read; full restart.

 Grid name : GLOBAL

 No water levels.
 No curents.
 Prescribed winds.

By changing DOMAIN%IOSTYP, I need to use the maximum memory on Hercules. However, the first time step's output of hs is still from cold start:

it should look like

Will you please suggest any further modifications I should try?

Thank you very much!

[aliabdolali]

Did you run the previous month with the same option?

[yunfangsun]

I didn't, I will try to run the previous month again.