[MITgcm-devel] bug-fix in seaice_init_fixed.F

[Martin Losch]

Hi there,

I just found and fixed a bug (and I also take full credit for creating this bug) in seaice_init_fixed.F that will affect all CS-integrations (actually all usecurviLinearGrid=.true. integrations) with seaice dynamics.
The coefficients for the metric terms k2AtC and k2AtZ were wrong because of one small typo (C->Z), so that k2AtC had the value of k2AtZ (so in effect a averaging interpolation error), but k2AtZ was zero. 

I suggest that you try if this fix (seaice_init_fixed.F r1.8 -> r1.9) dramatically changes your results. I don't expect any fundamental differences, because the metric terms are small anyway, but it's worth making sure ...

Martin