sattemp1979

Naval ports. Again, a miniscule fraction of the earth's surface. Coverage so sparse that it wouldn't register on a percentage scale with four places after the decimal point. You can't "accurately infill" data that hasn't been measured. What prevailed over far more than 99% of the earth's land and sea surface was unknown to any degree of accuracy before 1979; what prevailed in the depths of the massive heat sink called the ocean was not even partially addressed before ca. 2005. I'm not judging what you know about "S". My profession is both "T" and "E". I do suspect that you don't know much about "M".

Do you have any idea how much of the earth's land surface (to say nothing of its enormous oceanic heat sink) was covered by temperature observation before 1979? We're talking about a few times a day at a few hundred locations, over a few hundred years. Most of the temperature data before 1800 is from SE England; from 1900 onwards, the only better data are from the continental US. All amounting to a miniscule fraction of the data needed for a model...if the system were closed. It's not. No one has any idea what the sun (chief driver of earth's climate) will do next.

From the article: "The results: 10 of the model projections closely matched observations."

Not before 1979, they didn't. Because there isn't global temperature data before that. Nothing even close. There isn't ocean temperature data before this millennium. Not even remotely persuaded, I'm afraid.

I should really specify climate models in that bold claim. There's too little data (globally, from satellites, only since 1979) and global climate is not a closed system that can be experimented on. Nobody has any idea what the sun will do next; we can't even predict the purely terrestrial El Niño/La Niña trend past the notorious spring prediction barrier. Compare semiconductor models (which work very well) for example. Climate models are a joke. A bad one, at that.