... 1D

Note that it is not appropriate to directly compare the equivalent widths in Table 2 to estimate the net abundance difference between 1D and 3D as the 3D results are here only shown for one snapshot. The proper procedure is to add the 3D non-LTE corrections obtained here to the 3D LTE corrections published by Asplund et al. (1999) based on a long time-sequence with better numerical resolution.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.