My understanding is that Japanese does not distinguish direct and indirect questions. The grammatical model that I've found works best, is to treat か as a nominalizer. Given a (completed) predicate X, Xか functions like a noun meaning "the proposition that X is true", or "the question of X". (in this way, e.g. 問題 would bear the same relationship to interrogative か that 事 or やつ does to nominalizing の.)
Let's try applying that model. We have a main clause (for lack of a better term; the part following a conjunctive で)
その期間に氷床の大きさ (質量) がどれだけ変化したかの指標となる
Thus, その期間に氷床の大きさ (質量) fills the が position of the predicate どれだけ変化した ("it changed by some amount"). Adding か to that predicate nominalizes it: "the amount by which it changed", where "it" can be filled in with a translation of その期間に氷床の大きさ (質量) (left as an exercise).
Since we now have a noun, the subsequent の cannot be a nominalizer, but is instead the usual categorizing の. A "the amount by which it changed"-categorized 指標 ("indicator"), then, is an "indicator of how much it changed".
This then feeds to a quotative と and the verb なる. Thus, "Being a 涵養量と消耗量の差は氷床の「質量収支」, it became an indicator of how much that 期間に氷床の大きさ (質量) changed".
Unless the が-marked part actually belongs to なる, in which case I guess we end up with "Since it is a 涵養量と消耗量の差は氷床の「質量収支」, that 期間に氷床の大きさ (質量) became an indicator of the amount of change." But I think that would use に rather than the quotative と. That part is a bit above my level, it seems.