The "scientific consensus" on a given topic cannot be trusted. The very fact that it is appealed to as consensus, and not settled fact, should serve as a warning that weakens, rather than strengthens, an argument. What the consensus of experts does do is to set the Burden of Proof on the one who would either strengthen or challenge the consensus.
There is a fallacy hidden by the legitimate workings of the research process, making it difficult to spot: a combined Appeal to Authority and Appeal to Popularity. But a chain of two fallacies is still fallacious. The number and reputation of learned people who believe a proposition may affect how an idea spreads, but popularity does not affect truth or falsehood, particularly when the experts did not arrive at their opinions independently but were themselves relying on the authority of others in the field.
In a court of law, for instance, the truth is not determined by the number of experts one can line up to agree, else the better-funded or organized side would always win. It is up to the experts to present evidence, based on their experience and knowledge. The other side counters with expert evidence of its own, and the judge or jury evaluate. It is not the number of experts that decide the case, but the quality of the arguments each side can present. If a hundred experts declare that day is dark and night is light, the judges should find with one opposing expert noting the time and pointing to an open window.
Scientific Fact and the Zone of Conflict
Similarly, scientific belief is not, or should not be, formed on the basis of the number of scientists who believe one way or another.
All science is by definition unproven on some level; each fact is only accepted, not proved in the sense that a mathematician or logician proves something. For example, we don't know the speed of light, or that momentum is always conserved, and so on, only that according to observation and reasoning based on all of our observations to date, that the accepted fact or principle is reliable. Scientists speak of 'facts', but always with the knowledge that these 'facts' can be challenged. Challenging accepted facts requires supplying evidence and reasoning to counter them. It does not mean arranging a vote on the matter.
Experts on a given topic argue about it. The breadth of their discussion on a given topic varies, usually trending from the fundamental to the esoteric with the time and depth of study applied to the topic.
The number and reputation of scientists who accept something as probable or as fact only indicates the approximate parameters of the debate that surround a given idea. This can be thought of as the Zone of Conflict, the area for a given topic outside of which the facts are largely settled among experts. New evidence, or new arguments about old evidence, can change the popularity of a proposition, and thus expand or contract the Zone of Conflict in the field of study which pertains to that proposition, but they do not change its truth or falsehood.
Exaggerated Conflict
There is a fallacy lurking underneath the 'experts agree' layer and the 'scientific fact' layer, by which some seek to discredit a scientist or lay person based on how close his ideas are to Zone of Conflict. Conversely, some like to point out that no fact is ever totally proved, and thereby invent a controversy of fact when none really exists. Stating that all things are uncertain does not challenge a fact, any more than asserting a fact makes it so.
It is an Appeal to Ignorance to claim that because of the existence of the Zone of Conflict, any claim about that topic is as likely to be true as any other. But it is an Appeal to Authority and abuse of the Burden of Proof to claim that an argument is invalid because it is outside the Zone. In general, the range of discussion that takes place on a topic says nothing about the validity of a particular argument or the truth of a particular proposition.
Since the Burden of Proof is on the one making the assertion, it is the responsibility of the one who would expand the Zone of Conflict to show that a real question exists. Conversely, properly rebuffing such an attempt requires actually showing that the matter is settled and why the question is no longer open.
(Adapted from The Inescapable Logic of Global Warming)