As already noted by Dennis Schmoers, in most cases, the citation style will be dictated by the journal and the best way is to let a software deal with this for you, so you can switch to whatever the journal wants, once you choose or switch journals.
With that being said, your first choice is not very popular for the following reason: A numeric citation style (as opposed to citation by author name) has the advantage that you do not clutter the paper with citations. If you additionally order numeric references by appearance, you gain that when citing many references at once, they are usually together in the reference section (so the reader does not have to go back and forth or somehow memorise them when looking them up).
Compare the following:
Your first style combines the disadvantages of both other styles: You do not know who is cited, more space than necessary is occupied by citations, and when you actually want to look up those citations in the reference list, they are scattered all over the place.
If you impose alphabetical order on your numeric citations, the only advantage that you gain is that it is easier to see whether a certain paper is cited at all – if you know the name of the first author.
This is something I would not consider so important, and also if I want to know something like this, I usually have the paper on a computer, where I can do a full-text search – which has the great advantage of also finding cases where the author in question is not the first one.