restart

103 views
Skip to first unread message

chu yubin

unread,
Jun 4, 2022, 10:44:28 PMJun 4
to HYCOM.org Forum
Dear Alan,

I’m sorry to bother you. In order to more accurately simulate the global ocean in 2012, I want to convert 537_archm.2011_365_12.a of GLBb0.08 into restart.a and input it into GLBt0.72. Can this be achieved? If so, is it done using GLBt0.72/archive/archv2restart.csh?
Thanks for your patience.
Chu

Alan Wallcraft

unread,
Jun 6, 2022, 10:58:37 AMJun 6
to HYCOM.org Forum, chu yubin
Download


Run two scripts:

hycom/GLBb0.08/subregion/537_archm.2011_365_GLBt0.72.csh
hycom/GLBt0.72/archive/restart_537_2011_365.csh

The subregion script goes from GLBb0.08 to 3x coarser GLBB0.24 to 3x coarser GLBB0.72 and finally to GLBt0.72.

The restart script requires an existing "template" restart consistent with the archive file layer structure, you probably already have this.

The archive file had noisy interfaces, but I was able to run from the restart by reducing the time step from 1800 to 900 (this should only be required for the 1st model segment).

Alan.
Message has been deleted

chu yubin

unread,
Jun 9, 2022, 8:31:44 AMJun 9
to HYCOM.org Forum, alan.wa...@hycom.org, chu yubin

Dear Alan, 
     Thanks for your help. Because I don't have depth_GLBb0.08_11.[ab],i used depth_GLBb0.08_09m11.[ab]. And the script has generated a restart file for my model. But I have other questions:
      1. The GLBt0.72's regional. depth file is created by DBDB2. Will the script work if the regional. depth file is generated by GEBCO_2014_2D.nc?
      2. When I ran the HYCOM for 1 model year, the Kin energy was declining from 11500 to 4500. Is this correct?
      3.If I want to generate the restart.[ab] for other coarse resolution like 0.50 1.00 2.00,what should i do to create the 537_archm.2011_365_GLBaX.csh and restart_537_2011_365.csh?
      Thanks for your patience.
      Chu

Alan Wallcraft

unread,
Jun 13, 2022, 3:55:44 PMJun 13
to HYCOM.org Forum, chu yubin, Alan Wallcraft

1. Yes, any valid bathymetry will work.  Note that mismatches between the source and target bathymetry can lead to anomalies near the bottom but these are usually harmless.

2. I saw the same thing from my 1-month run:

gaffney01 2485> grep -h "region-wide mean Kin. Energy" 537y015A.log
   490945 (1915/001 00) region-wide mean Kin. Energy:    11935.8313999618
   491328 (1915/005 00) region-wide mean Kin. Energy:    11882.7658145945
   491904 (1915/011 00) region-wide mean Kin. Energy:     8662.4388646673
   492480 (1915/017 00) region-wide mean Kin. Energy:     7544.1197176876
   493056 (1915/023 00) region-wide mean Kin. Energy:     6913.4861341761
   493632 (1915/029 00) region-wide mean Kin. Energy:     6384.0411988944
   493920 (1915/032 00) region-wide mean Kin. Energy:     6212.8082575097

4500 is about what we get for GLBt0.72 cases without assimilation.  The surprising value is 11500, but that is likely from higher initial velocities from 53.X.  The two 3x averaging steps where supposed to smooth these out.

3.  For resolutions "close" to 0.72 (e.g. 0.5 and 1.0) all you need to do is generate a regional.gmapi_GLBB0.72 for your region and replace the last step with GLBB0.72 to newregion.  This would be done with a gmapi.csh script in hycom/GLBB0.72/subregion,  see for example https://github.com/HYCOM/HYCOM-examples/blob/master/GLBb0.08/subregion/gmapi.csh  For even coarser resolutions, you could add let another 3x step from GLBB0.72 to GLBB2.12 and then on to your final resolution.

Alan.

chu yubin

unread,
Jun 15, 2022, 3:19:15 AMJun 15
to HYCOM.org Forum, alan.wa...@hycom.org, chu yubin
Dear Alan,
Thanks for your advice. I have tried to  configure the resolution to 0.50, but the  procedure is killed when I run the  restart_537_2011_365.csh . Could you please tell me how to resolve this error?  All files can be downloaded here.(restart_out.[ab] is  "template" restart)

12973761 mean basin temp, saln, dens        NaN       NaN       NaN
 12973761 (2012/001 00) mean      SSH (mm):     NaN  (-8.4E+03 to  3.6E+04)
 12973761 (2012/001 00) mean MontgPot (mm):     NaN
 
 error - NaN or Inf detected
 
 12973761 (2012/001 00) mean BFL (m^2/s^3):     NaN hfl:     NaN
 12973761 (2012/001 00) mean HFLUX (w/m^2):     NaN sst:    0.00  ml:     NaN
 12973761 (2012/001 00) mean WFLUX (mm/wk):     NaN sss:     NaN
 12973761 (2012/001 00) mean  ice thk. (m):********  temp:-18.839 pcen: 65.351
 12973761 (2012/001 00) mean SH I thk. (m):********  temp:-36.148 pcen: 24.702
 12973761 (2012/001 00) mean NH I thk. (m):********  temp: -8.321 pcen: 40.649
 12973761 (2012/001 00) mean mixl thk. (m):     NaN  temp:    NaN saln:    NaN
 12973761 (2012/001 00) mean surf thk. (m):    3.00   sst:    NaN  sss:    NaN
 
 error - NaN or Inf detected
 
 12973761 (2012/001 00) mean clim thk. (m):   19.20   sst: 19.353  sss: 34.732
 12973761 (2012/001 00) mean L  1 thk. (m):    0.30  temp:    NaN saln:    NaN
 12973761 (2012/001 00) mean L  2 thk. (m):    0.00  temp:  0.000 saln:  0.000


Thanks for your patience.
Chu

Alan Wallcraft

unread,
Jun 17, 2022, 2:19:26 PMJun 17
to HYCOM.org Forum, chu yubin, Alan Wallcraft
The download link needs access permission, I requested access but with no response.  Unless this needs to be restricted, the easiest fix is to set "anyone with the link" permission.

Please confirm:

a) The run from climatology to create the template restart worked.
b) This run has half the baclin time step of the run from climatology.

An initial NaN can be from NaN in the input files, but it is also often from batrop being too large.


Typically we use about 75% of the minimum CFL:

COAPS 6268> cat depth_GLBt0.72_15_batrop.b
bathymetery from 2-minute NAVO/NRL DBDB2 global dataset; NGDC World Coast Line
i/jdm =  500  382                                                              
plon,plat range =  -286.19775   433.79773   -77.94000    89.80036              
landmask Black/Red/Baltic Seas, other narrow inlets; B-grid (CICE) fixes      
Merged with depth_GLBt0.72_87 coastline in Antractic; 1st row all land        
gsp:  min,max =     14.004   283.508
cfl:  min,max =     90.773  3949.153

We are using batrop=56.25.

In general, you want this as large as possible and if it is too large the model will immediately blow up (so it is easy to set the right values).

Alan.

chu yubin

unread,
Jun 18, 2022, 7:46:19 AMJun 18
to HYCOM.org Forum, alan.wa...@hycom.org, chu yubin
Dear Alan,
I'm sorry about the file, I changed the permissions of the file and  added the new files. For the 0.5 resolution experiment, I previously set baclin = 540 and batrop = 33.75 based on the cfl calculated by batrop.csh, and also set wndflg = 0 to generate restart_out.[ab]. After that I set baclin to 270 and batrop to 16.875, but the problem still occurs.I think there is a problem with restart_in.[ab], but I don't know how to fix it. Could you please help me to check it?
Thanks for your patience.  
CHU

Alan Wallcraft

unread,
Jun 29, 2022, 1:12:51 PMJun 29
to HYCOM.org Forum, chu yubin, Alan Wallcraft
Your gmapi file is wrong.

 input domain assumed to be periodic
output domain assumed to be non-periodic
 input domain is curvilinear
output domain is curvilinear

GLBB  (500x367) to GLBa0.50 (maxinc=5)
xmap:  min,max =    125.7619    376.0000
ymap:  min,max =    336.6857    367.0000

The output domain  should be periodic. 

I reran your script using ~/HYCOM-tools/subregion/src/isuba_gmapi and it worked for me:

COAPS 4525> cat regional.gmapi_GLBB0.72.b
GLBB0.72 (500x367) to GLBA0.50 (maxinc=99)
xmap:  min,max =      1.0000    500.4943
ymap:  min,max =     44.7804    299.6039

Try again using the HYCOM-tools isuba_gmapi.

Alan.
Reply all
Reply to author
Forward
0 new messages