[Rebelgecko]

I recently read a very convincing (to me at least) debunking of that first link, but unfortunately I'm not able to find it right now. But from memory:

The claim is that the temperature data in the first map is made up because some of its data isn't present in the second map (just look at Africa, South America, and the oceans). HOWEVER:

The first map shows a combination of two data sets. The second map only shows one of the data sets used in the first map. Not surprisingly, this means that the second map has less data.

The two graphs are also covering different time periods. The second graph covers a smaller time period, so presumably that lets them be more specific about where the temperatures are recorded. The ERSST data in the first graph goes back more than 100 years, so they don't have quite the same precision and consistency in terms of where the measurements came from. The bigger grid squares in the first map reflect that some of the data is interpolated as weather stations moved from city to city, new weather stations start coming into play halfway through the dataset, etc.

EDIT: Apparently the debunking I read was on Stackoverflow of all places. I must've fallen down that rabbit hole while looking up something totally unrelated. Here's the link: http://skeptics.stackexchange.com/questions/37119/did-noaa-p...

[cool_look]

From the stackoverflow comments its clear that theoretical models have been proposed to infer the temperature on land from the ocean database.

the main concern is that the models are pure junk science.

so saying "our models say it is so" is precisely the problem

[Rebelgecko]

Are there any specific concerns with the methodology? I didn't see any mention of that in the first link.