combo of LMD and WET_DRY blows up

[kate]

Question about this code in step2d:

# if defined WET_DRY && defined MASKING

zeta(i,j,knew)=zeta(i,j,knew)+ &

& (Dcrit(ng)-h(i,j))*(1.0_r8-rmask(i,j))

# endif

Do you really want to do this inside the land mask or just inside the rmask_wet? It causes trouble in the LMD_SKPP inside the land mask when h is big (say 312m). It leads to a wacky computation of the light at the bottom of the mixed layer if zeta=-312. If this is indeed correct, I would advocate setting h to hmin inside the land mask after it is read in.