ROMS wet/dry mask

[lanerolle]

I am running ROMS for AK using wetting-drying and as my run is quite long and I need to use a small time-step, I am doing it in segments using the PERFECT_RESTART CPP option. For non-wetting-drying applications, this CPP option has worked very well.

Now, when a ROMS run is initialized, the field wet_dry_masking (ie. the wetting-drying mask) is all zeros - ie. its dry everywhere and as the run progresses, certain places begin to dry and wet but all of the locations given by mask_rho/mask_u/mask_v remain permanently dry as expect.

I however find that when ROMS does a perfect restart, rather than continuing to evolve the wet_dry_masking from where it stopped at the previous run/segment, the run again begins by setting it all to zeros everywhere.

Is this a bug or was the algorithm designed to do so? If I did a run with 10 run segments (from perfect restarts), can I trust the time-evolution of wet_dry_masking for the total number of segments or should I assume that the wetting-drying within each segment is trustworthy and that the previous/next segment is a totally differnt story?

[kate]

It sounds like you might have to debug PERFECT_RESTART for the wet-dry case. Is it giving the same answer with restart as with running through?

[arango]

Yeah, it is possible that we need to initialize the correct wet/dry masking during restart. I will make a note about this and check when I get the time. Now, I am busy with the ROMS workshop.

[turuncu]

Hi,

I have same issue and i just wonder that is this bug fixed or still open? I am using ROMS 3.5 with ice model and it seems that it has still problem. When i try to restart the model with perfect restart option, the wet-dry mask is not correct at the beginning. i think that the problem is in the Nonlinear/ini_fields.F because it does not initialise the wet-dry masking correctly. It just checks the critical depth criteria but in fact that the correct wet-dry mask is in the restart file itself. Anyway, if it is not fixed yet, could you tell me the correct place that i need to modify. I think that it must be done in Nonlinear/ini_fields.F like other variables such as zeta, ubar, vbar etc. but i am not sure.

Regards,

--ufuk

[kate]

I believe some patches involving the wet-dry mask have come out in the past three years. I have not tried WET_DRY with the ice model in a long time, though, and only recently got PERFECT_RESTART working with the ice at all. I suggest you update and see what needs doing on the most recent code (though I did stop short of the patch taking it to ROMS 3.7 which has nothing to do with this particular issue).

[turuncu]

Thanks Kate. I start to work in this problem. I hope i'll fixed it soon. By the way, i am using ice branch which was retrived in 20-03-2012 (i think that it is based on ROMS 3.5) and i just wonder that did you do any major code change after that. I am getting ice frazzle problem somtimes when i use wet-dry with ice model.

Thans again,

--ufuk

[kate]

Your code is a year old. In that time I have fixed the ice PERFECT_RESTART and fixed the frazil ice trouble to make it die instead of endlessly complaining. Trouble in frazil ice could be due to negative water depths or to bizarre values of T and S (I've seen both). I doubt I've fixed the source of your troubles.