saving average and diagnostic output when using restart

General scientific issues regarding ROMS

Moderators: arango, robertson

Post Reply
Message
Author
jxiong
Posts: 4
Joined: Wed Feb 10, 2021 8:06 pm
Location: University of Washington

saving average and diagnostic output when using restart

#1 Unread post by jxiong »

Dear ROMS users,

I wonder if anyone has experience saving average and diagnostic output when initializing a run from a restart file. Is this functionality supported in ROMS, or are there known issues when attempting this?

In a 1-day test run with ROMS/TOMS version 4.2, I'd like to output daily-averaged average and dignostic variables. I defined the following:
• PERFECT_RESTART, AVERAGES, and DIAGNOSTIC_TS,
• NTIMES = 2160, DT = 40
• NAVG = 2160, NDEFAVG = 2160
• NDIA = 2160, NDEFDIA = 2160
• NRREC = -1, ININAME = '../ocean_rst.nc' (the time is 2024-10-01 00:00:00)
• DSTART = 19997, and TIME_REF = 19700101.

However, the model crashed with an error: "unabel to open diagnostic netCDF file: ../ocean_dia_0000.nc".

After searching the forum, I found a potential solution in this post: viewtopic.php?t=5373 that suggested changing DSTART to a previous date. I tried DSTART - 1 (day) and now the model can save oean_dia_0002.nc without errors. Yet, I'm conerned about the DSTART adjustment. While this shift avoids the issue with the _0000 file name, I noticed that it appears to affect the output, for example, salinity in the history files changes when using DSTART - 1. Could this shift actually cause time misalignment such as in the forcings?

Another interesting observation is that regardless of how much earlier I set DSTART (e.g., DSTART - 1 vs. DSTART - 10), the diagnostic, average, and history variables remain identical, with only the output filename changed (e.g., ocean_dia_0002.nc vs. ocean_dia_0011.nc). That makes me wonder whether DSTART is primarily influencing the file indexing...

Any thoughts or suggestions would be greatly appreciated.

Thanks,
Jilian

Post Reply