"Most cited papers" in a given field/period is fairly easy through something like Web of Knowledge; it's a few clicks to find out that both Scopus and Web of Knowledge list this as the most cited paper in chemistry in 2014, for example. (Tellingly, they only agree on three of the top five...)
Doing this for authors is more challenging, though - the Google Scholar registered authors method suggested above is effectively an opt-in database and so there's no way of telling who's not in it. Scopus has fairly good author-citation indexing but I'm not sure if it's set up to answer this specific question.
A couple of studies with similar questions:
Abramo et al looks at a defined set of researchers (all Italian academics) and shows a way to identify the most highly cited authors within them - basically just WoK citation counts. Robust but tedious, and you probably don't want to start with a list of "all astrophysicists"...
Another approach is to pick the most highly cited articles in your field (assuming that the bulk of citations are in these, not the long tail) and look at the authors of those papers. See, eg, Choi et al, which is "the most cited authors of the 250 top papers in the field", or Uthman et al, which looks at the authors of the top hundred medical reviews.
You could combine these - pull out, say, the fifty most cited papers, look at their authors, and do full citation counts on these using WoK. It seems likely that the most cited authors overall will be an author on one or more of the top papers.