IANAClimatologist. This[1] chart is not subject to the "missing data" problem; as cited by wikipedia it is composed from freely available data. In addition by including satellite data that is free of an entire class of measurement artifacts it substantially limits the plausible error alleged by the article. The best interpretation of the article is that it is an astonishingly narrow picture of a small slice of an enormous field.
Isn't satellite data subject to a different class of measurement artifacts? Since it is measuring reflection at potentially different layers of the atmosphere, some high, some low?
The noise is different but the signals are strikingly similar. That's a strong point in favor of the model IMO, but as I said this is a complex, subtle field in which I'm not an expert. Imagine how badly the popular media gets computer stories and think that we're making at least as bad a hash out of this whole field.
[1]http://en.wikipedia.org/wiki/File:Satellite_Temperatures.png