Since 2004, when storm waters introduced Pacific carp to Hoffman Lake,

Okay, let's take this baby down. Many of the responses in this forum focus on isolated pieces of this question, but I think it might be beneficial to do a full analysis, focusing on the critical-thinking strategies that unlock problems like these. As you study for the GMAT, it is very beneficial to identify patterns and strategies useful for multiple questions, not just one. Here is the full "GMAT Jujitsu" for this question:

Our first item of business is to recognize the problem type. Identifying exactly what the problem is asking is one of the most "critical" things you can do with critical reasoning questions. The question stems asks us to identify a statement that "must be true" given the information in the stimulus. At the basic level, this is an "Inference" question type. Inference questions have a very high burden of proof, as indicated by the phrase "must be true." If an answer either contains new information or leaves open possibilities so the statement isn't necessarily true, we can eliminate it. In my classes, I call these two measuring sticks the "No-New-Information Filter" and "No Wiggle Room." Let's analyze each answer choice to see how it measures up against these standards.
If you picked this answer, you likely made the assumption that the introduction of Pacific carp into the lake caused the population of smallmouth bass and sand trout to decrease. But this is a "Correlation is not Causation" fallacy. The passage never says that the introduction of Pacific carp caused anything, and the passage also never suggests what will happen if the Pacific carp are removed. Answer choice A definitely fails the "No-New-Information Filter." Eliminate it.
The passage states that currently "Hoffman Lake is home to more sand trout than any other type of fish." However, we don't know if this is the case in 2004. I have seen some people do the full mathematical analysis on this answer using the rates of decrease quoted in the stimulus, but such an analysis is completely unnecessary. Don't fall for such a time-wasting trap. After all, the rates mentioned in the problem only describe the decreases in the populations of sand trout and smallmouth bass. While we might be able to numerically prove whether sand trout was more numerous than smallmouth bass in 2004, proving that sand trout was more numerous than "any other type of fish" is way beyond the scope of the problem. The problem never mentions how other fish populations have changed since 2004. Answer choice B leaves open the possibility that another unmentioned fish species may have been dominant in 2004. Just because a passage doesn't mention other species doesn't mean that other species can't or don't exist. I call this the "Lack of Evidence is Not Evidence" fallacy. Because we cannot prove that this answer "must be true," we can eliminate it. Answer choice B has too much wiggle room.
Answer choice C is easy to eliminate. It contains new information. There is nothing in the stimulus that indicates the Pacific carp must be predators. If you assumed this, you might have fallen for the "Correlation is not Causation" trap, very similar to the one found in answer choice A.
Note that the stimulus states, "Hoffman Lake is home to more sand trout than to any other type of fish." This fact is very clear in the argument. "Any other type of fish" naturally includes smallmouth bass, and the "some point" when sand trout are more numerous than any other type of fish is right now. There is no wiggle room here. Answer choice D must be true. It is the correct answer.
Answer choice E has the same problem that answer choice B does. Besides the fact that Pacific carp were "introduced" to Hoffman Lake, we are given no information on the size of this population. We also have no idea if another unmentioned fish species could be dominant in the future. This answer fails both the "No-New-Information Filter" and the "No Wiggle Room" test. Eliminate it.

In the end, only one answer choice "must be true" given the facts in the stimulus. All other answer choices either contain new information or leave some "wiggle room." The strategy here is clear: focus on the specific wording in the question stem and don't add in extra information or assume causal relationships that go beyond what the problem explicitly states. Problems like this can be quite easy if you leverage the structure of the question to your advantage.