For two way nesting model with wet-Dry on, if the fine grid treat a area as ‘dry’, it will pass a zero value to the coarse grid, even this area is considered as 'wet' in coarse grid. This somehow did not affect too much the coarse model simulation, but will leave some unpleasant dots on temperature plots. It seems tricky to treat these areas with different mask in different grid. I think it would be better to update the coarse model mask (within the fine model domain) with fine model mask for each time step .
See the code below. The highlighted line will compute an zero value to my_sum where Fmsk=0;This is OK to velocity, and not OK to temperature and salinity.
my_area=dxF(i+Iadd,j+Jadd)*dyF(i+Iadd,j+Jadd)
my_areasum=my_areasum+my_area
# ifdef MASKING
my_sum=my_sum+ &
& F(i+Iadd,j+Jadd,k)*my_area* &
&
MIN(1.0_r8,Fmsk(i+Iadd,j+Jadd))
my_count=my_count+MIN(1.0_r8,Fmsk(i+Iadd,j+Jadd))
# else
my_sum=my_sum+ &
& F(i+Iadd,j+Jadd,k)*my_area
# endif