Has NTIMES meaning changed?

[mmagaldi]

Hello ROMS users,

I recently got back to ROMS to set up a new configuration after almost 3 years, downloading the version #854 in July this year. The previous version I used was #736, downloaded in June 2014.

I am puzzled by what I should insert as a value in NTIMES in the new version.

In the past, it has always been clear to me that NTIMES determines the timestep on which the run will end. I guess the reasons behind this choice are still listed in the official ROMS Wiki page at https://www.myroms.org/wiki/ntimes.

In the new version NTIMES seems instead to play a different role. For example in the new get_state.F theres is a new line

Code: Select all

ntend(ng)=ntstart(ng)+ntimes(ng)-1

where it is added to the number of timesteps set by the initial time (ntstart) while ntend determines now the timestep on which the run will end. Note that I have always used (and double-checked that) dstart=0.

If I tracked this correctly, the changes are still present in the current trunk (#870) and were introduced with the following ticket https://www.myroms.org/projects/src/ticket/735 where "a correction with the management of ntend(ng)" is cited.

Is NTIMES still the timestep on which the run will end? Sorry but I could not find a message in the ROMS list about the new meaning of NTIMES and I am wondering whether or not I am getting this right.

Looking forward to your replies,

thanks in advance,

m.

[kate]

It did indeed change - in fact it changed to the meaning it had back years before ROMS. It now means the number of times to be run in the current run, so if you want to run in chunks of 100 days, you don't have to change NTIMES after every restart. However, if the model crashed at day 92, to get back on track you need to set NTIMES for 108 days just the once. Whatever, but the wiki should be updated.