lyfan512For questions like this, it's important to recognize the underlying ground rules. Regardless of the question, we never question the premises (the stated facts) of the argument. Rather, we look for the disconnect between these premises and the stated conclusion. If we want to weaken a conclusion, we knock down its support not by denying the premises, but by showing how the premises may not lead to the conclusion as stated. I agree that this question is worded more confusingly than usual, but that is still our task. If the premises don't lead to the conclusion, then the conclusion lacks support!
Here, the argument is using a correlation (droughts, collapse) to attempt to make a very specific causal claim: it's not just that droughts cause collapse, but they cause it by leaving societies without the needed resources. That's where D fits in. If these societies had plenty of resources stockpiled at the time of collapse, then it's not at all clear that drought had this particular effect. So while the premise may still be true, it doesn't clearly support the conclusion the author wants. That's the only sense in which we have "weakened the support."
Another important point is that even if we did want to weaken a premise, answer choice A doesn't do that. In fact, we'll never see an answer that attacks a premise directly (although it may shed more light on one), since both the premises and the answer choices (for a str/wkn) are taken as true. In that way, these are like Data Sufficiency. Statements 1 & 2 can't contradict each other, since they are both true! So why isn't A knocking out a premise? Because "some do" is never negated by "some don't." If I say "Some people won't rest until they get an 800, so I can create a service specifically for the most ambitious students," that underlying premise would not be negated by "Some people would be happy with 750, 700, or even 500." This doesn't make the previous statement any less true. Subbing "many" or "often" for "some" makes no difference either, since these are vague terms. Basically, the author is saying that there were a number of times when droughts lined up with collapses. This doesn't mean they always lined up, and the argument doesn't rely on that, so A doesn't really change the status of the argument in any way.
I hope that helps!