Re: [swanmodel-users] Long Simulations with Multiple Output Files

[gustavo s. c. <gus...@ho...>]

Hello Joao,

We had already talked once about this file size limit.
So, okay, I will do annual simulations and increase the parallelism.

Thank you very much for the answer, João!

Gustavo Correia.


______________________________________________________________________________________

Gustavo Souza Correia

Mestrando em Geociências pela Universidade Federal do Rio Grande do Sul - UFRGS

Oceanógrafo pela Universidade Federal do Maranhão - UFMA

Técnico em Meteorologia pela Universidade Estadual do Maranhão - UEMA

Currículo lattes CNPQ: http://lattes.cnpq.br/5164631760543791



________________________________
De: João Albuquerque <ja...@au...>
Enviado: domingo, 31 de maio de 2020 22:04
Cc: swa...@li... <swa...@li...>
Assunto: Re: [swanmodel-users] Long Simulations with Multiple Output Files

Hello Gustavo,

It is possible to set more than one output per variable, however, as far as I know SWAN won't stop adding data to the former one. For example, if you set this output:

BLOck 'COMPGRID' NOHEAD 'Hs_era5_2015.mat' LAY 4 HSIGN OUTput 20150101.000000 60 MIN
BLOck 'COMPGRID' NOHEAD 'Hs_era5_2016.mat' LAY 4 HSIGN OUTput 20160101.000000 60 MIN

SWAN will start writing the file 'Hs_era5_2016.mat' at the first of January of 2016, but it will continue to add data to your 2015 file. In this case, depending on your grid's resolution the model will crash whenever your .mat file reaches the limit size of 4GB (or 8GB, not sure right now).

If you are running SWAN on a cluster I'd recommend you to run SWAN in yearly or 6-monthly simulations. In this case, instead of using hotfiles, which would tie you to one serial run, you can increase your parallelism by warming up your grid with about 36 hours (please double check whether this interval works for you as your mileage may vary) of previous boundary and wind forcings before the first output cycle.

Good luck,
João

On Mon, 1 Jun 2020 at 10:39, gustavo souza correia <gus...@ho...<mailto:gus...@ho...>> wrote:
Dear Colleagues,

I'm working on a hindcast of several years and I want to know if it is possible to make the SWAN write more than one output per variable, that is, in addition, write each variable from year to year, without having to do several simulations.

Currently this part of the script looks like this, but I want to let the model run for more years in order to generate annual outputs:

$

 BLOck 'COMPGRID' NOHEAD 'Hs_era5_2015.mat' LAY 4 HSIGN OUTput 20150101.000000 60 MIN
 BLOck 'COMPGRID' NOHEAD 'Dm_era5_2015.mat' LAY 4 DIR OUTput 20150101.000000 60 MIN
 BLOck 'COMPGRID' NOHEAD 'Dp_era5_2015.mat' LAY 4 PDIR OUTput 20150101.000000 60 MIN
 BLOck 'COMPGRID' NOHEAD 'rTp_era5_2015.mat' LAY 4 RTP OUTput 20150101.000000 60 MIN
 BLOck 'COMPGRID' NOHEAD 'Tm01_era5_2015.mat' LAY 4 TM01 OUTput 20150101.000000 60 MIN
 BLOck 'COMPGRID' NOHEAD 'sTp_era5_2015.mat' LAY 4 TPS OUTput 20150101.000000 60 MIN
 BLOck 'COMPGRID' NOHEAD 'wind_era5_2015.mat' LAY 4 WIND OUTput 20150101.000000 60 MIN
 BLOck 'COMPGRID' NOHEAD 'grid.mat' LAY 4  XP YP
 $
 $
 $
 $
 TEST  ITEST= 1     ITRACE= 1
 COMPUTE NONSTAT 20150101.000000 15.00 MIN 20151231.230000
 HOTF 'hot_1_20151231.230000'  FREE
 STOP


Thank you in advance!

Gustavo Correia.


______________________________________________________________________________________

Gustavo Souza Correia

Mestrando em Geociências pela Universidade Federal do Rio Grande do Sul - UFRGS

Oceanógrafo pela Universidade Federal do Maranhão - UFMA

Técnico em Meteorologia pela Universidade Estadual do Maranhão - UEMA

Currículo lattes CNPQ: http://lattes.cnpq.br/5164631760543791



_______________________________________________
swanmodel-users mailing list
swa...@li...<mailto:swa...@li...>
https://lists.sourceforge.net/lists/listinfo/swanmodel-users


--
João Claudio Albuquerque
PhD. Candidate, School of Environment, The University of Auckland,
Level 4, Room 449, Science Centre, 23 Symonds Street, Auckland, NZ.
+64 021 02365705