#2106 closed Bug (wontfix)
Floating invalid in limthd_dh.f90
| Reported by: | tian | Owned by: | vancop |
|---|---|---|---|
| Priority: | low | Milestone: | |
| Component: | LIM3 | Version: | v3.6 |
| Severity: | major | Keywords: | LIM v3.6 |
| Cc: | uwe.fladrich@… |
Description
The problem is that I get a error of floating invalid in limthd_dh.f90, line 460 & line 487. It happens in the coupled configuration (ifs_grid=T255L91, nem_grid=OR
CA1L75) when I apply a sea ice restart file, remapped from ORCA025L75 to ORCA1L71. The coupled model crashes in the first time-step.
I detect some grid points having zero in the denominator of a fraction " ( ztmelts - rt0 ) / ( zt_i_new - rt0 ) ", where ztmelts, rt0 and zt_i_new all equal to 273.15. It is because regridding introduces zero-value to those grid points. I solve it by adding epsi20 to the denominator. As the numerator of the fraction is also zero, the fraction will be zero with the new denominator ( zt_i_new - rt0 + epsi20).
Commit History (1)
| Changeset | Author | Time | ChangeLog |
|---|---|---|---|
| 9883 | clem | 2018-07-05T17:57:31+02:00 | solve issue raised in ticket #2107 and maybe in ticket #2106 |
Change History (6)
comment:1 Changed 5 years ago by clevy
- Owner set to vancop
- Status changed from new to assigned
comment:2 Changed 5 years ago by clem
comment:3 Changed 5 years ago by clem
In 9883:
comment:4 Changed 5 years ago by clem
I changed a couple of things in the ice code that could produce this error. If you take r9883, you should not need the epsi20 trick you used, and the solution should be cleaner.
comment:5 Changed 4 years ago by vancop
- Resolution set to wontfix
- Status changed from assigned to closed
obsolete for nemo v4
comment:6 Changed 16 months ago by nemo
- Keywords v3.6 added
When zt_i_new = 273.15 it usually means that the sea surface salinity = 0 at this point.
Therefore I would suggest to check the surface salinity first.