My understanding is that the Antarctic still remains the source of a lot of uncertainty in climate models. This paper, for instance, is very interesting, from James Hansen et al:
