question about ice frazil ...

Discussion about modeling ice with ROMS

Moderators: arango, robertson

Post Reply
Posts: 128
Joined: Tue Feb 01, 2005 8:21 pm
Location: Istanbul Technical University (ITU)

question about ice frazil ...

#1 Post by turuncu »


I am running ROMS-ICE and after 9 year simulation, i got the "trouble in ice_frazil" error than model crashed. The main thing is that i am also using wetting/drying with ice. I also checked the restart file (one step before blow-up) and some single grid points behave so differently than the grid cells near it (they are warmer etc.). So, i just want to know that is there anyway to overcome this problem. It could be caused by the wetting and drying but i am not sure.

The masking of wetting and drying also little bit strange. In my example, my minimum depth is 5 meter and in the middle of shallow (or almost middle) region (north caspian sea) the WETDRY_MASK_RHO value is zero (so it is shown as land). Maybe i am wrong but i expect that the drying start from the coast and move towards the center of the domain. So, it seems that the model creates artificial lands in part of my domain which is in the same depth with the neighbor cells. This could be issue of water balance because the wetting and dying compare the depth+zeta with dcrit so if the zeta is a large negative value such as -2 meter or more then the actual depth of the grid cell will be shallower and then it violates the dcrit threshold and ROMS block that cell. The water balance becomes an important issue when you consider to do relative long run in a closed basin like a lake. I also set the EMINUSP_SSH together with EMINUSP to prevent the change in the zeta but it does not help.

Any suggestion will be helpful for me at this point.



User avatar
Posts: 3955
Joined: Wed Jul 02, 2003 5:29 pm
Location: CFOS/UAF, USA

Re: question about ice frazil ...

#2 Post by kate »

I don't know what your exact problem is, but I can tell you that the print statements in ice_frazil are like a canary in a coal mine, warning of many kinds of problems. Some of those problems shouldn't be happening with WET_DRY defined. It looks like you have the right approach for getting in and looking at things right before it goes bad.

Post Reply