To me, scientific evidence is not "a whole bunch of scientists think X." To me, scientific evidence is "Here is an experiment or a naturally-occurring event where the results are extremely unlikely to occur unless X is true." That is what I find missing in the man-made global warming story.
I doubt that people claim that consensus is the same thing as scientific evidence. Instead, they rely on consensus because (a) It strongly correlates with scientific evidence, and (b) It is much easier for non-experts to check for consensus than to directly review the scientific evidence.
In any case, isn't Arnold's standard of scientific evidence strangely strict? Few results in economics rest on an "experiment or a naturally-occurring event where the results are extremely unlikely to occur unless X is true." The same goes for evolutionary biology. Some will be delighted to dismiss both as pseudo-sciences, but the real lesson is that Arnold's standard is excessive. Contrary to his account, the typical scientific conclusion rests on large body of diverse pieces of evidence. Most of these pieces of evidence are not, by themselves, very convincing; it is the totality of the evidence, not any particular finding, that provides the basis for scientific judgment.