getting different results with redo of old run

Report or discuss software problems and other woes

Moderators: arango, robertson

Post Reply
Message
Author
wendy
Posts: 15
Joined: Thu Jan 06, 2005 5:19 pm
Location: Institute of Ocean Sciences

getting different results with redo of old run

#1 Unread post by wendy »

I recently redid an old run, this time with diagnostics turned on. I output the history files as well, just to be sure I haven't changed anything in the meantime and have found that the results are slightly different.

From looking at the log files, the only difference between the old run and the new one is the value of the Haney number and the initial basin volumes. I output grid parameters to my history files, so I was able to compare h and Cs_r from the two runs. Both parameters were exactly the same between the runs, as was the land mask and the initial zeta, so how is it possible that the Haney number or the basin volumes could be different?

Is there something else I should be looking at?

Thanks very much,
Wendy

User avatar
kate
Posts: 4091
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: getting different results with redo of old run

#2 Unread post by kate »

You could check out an old version of ROMS to see if the computation of the Haney number or the volume has changed. Are the results otherwise the same?

wendy
Posts: 15
Joined: Thu Jan 06, 2005 5:19 pm
Location: Institute of Ocean Sciences

Re: getting different results with redo of old run

#3 Unread post by wendy »

I just realized I wasn't completely clear in my first post. I meant the only difference in the log headers are the Haney number and the basin volumes. The history file output is definitely different between the two runs, otherwise I wouldn't bother with checking into this. After thirty model days the solutions are broadly similar, but features are different enough that I'm getting a change in surface u of up to 30cm/s (which in a region of max u ~= 50cm/s is significant).

I don't *think* I have updated roms between the two runs, but it's possible. Is there a way to check?

Thanks!

User avatar
kate
Posts: 4091
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: getting different results with redo of old run

#4 Unread post by kate »

New ROMS on a system with svn working reports something like:

Code: Select all

 SVN Root URL  : https://www.myroms.org/svn/omlab/branches/kate
 SVN Revision  : 636:641M
near the top. I'm not exactly sure what that 636:641M thing means, though. Do you have that and do they match in the two runs?

wendy
Posts: 15
Joined: Thu Jan 06, 2005 5:19 pm
Location: Institute of Ocean Sciences

Re: getting different results with redo of old run

#5 Unread post by wendy »

doesn't look like I've updated. Both log files say:

SVN Root URL : https://www.myroms.org/svn/src/trunk
SVN Revision : 154M


Just to make sure I wasn't crazy, I just tried doing two identical runs (different history file output name and different log file name, but otherwise the same) and the history file output was identical, but the Haney numbers were not. 6.887 vs 6.925, so there's something weird going on with the Haney number calculation, but it looks like that's not what's causing the differences in my runs. I'll keep searching.

Post Reply