is a lot easier to identify then temperature within a tenth of a degree a year when 2 of the 7 continents don't even have a temperature record.
If you are looking at 20,000 years and 400 feet it doesn't make much difference if the start date is off by a thousand years or the initial sea level is off by 20 feet. It's still about 2 feet per century which is double the current rate.
If you are determining the temperature of all of Africa for a specific month by one thermometer in Sao Tome and another in Algeria it makes a hell of a difference if:
1. The thermometers are accurate.
2. The people who recorded the data did it accurately.
3. You have the gaul to claim that you can do it within a tenth of a degree.
Since you asked I'll be happy to justify the use of satellite data. I use satellite data because the surface data is crap. It's crap today, it will be crap tomorrow and it was probably crap 50 years ago. The reason it is crap today is that the people responsible for the integrity of the system don't give a rats ass about quality control. Yes I know "I'm an idiot". It doesn't matter to you that 90% of the USHCN stations are out of compliance with their own requirements because "deniers" identified them as such. The fact that they documented their findings is unimportant to you. 58% are inaccurate by 2 degrees or more and another 11% are inaccurate by at least 5 degrees. None of that matters to you because you can type insults.
Joe Romm dismisses it as "propaganda". Notice how he doesn't dispute it. He just dismisses it. I've seen it explained that the problems can be "adjusted" away but they didn't even know which stations had biases. How do you adjust something that you haven't even identified? Speaking of identifying, did you know that it got down to -27 degrees in Key West last Saturday. It's a little joke in the "denier" community. We're waiting to see how long it will take the high quality NOAA people to notice. Some how a idiotic "denier" who doesn't get paid spotted it but the highly trained scientists didn't. Maybe they should write a line of code saying that if a temperature is off by more then 20 degrees C. (36 F.) from expected it should be checked out. Key West was off by about 107 degrees.
http://www.srh.noaa.gov/data/obhistory/KEYW.htmlGISS shows the temperature of Kazakhstan centered around 46 degrees East and 57 degrees North as 19.4967 C (35 F)above normal in January of 2002. Since it was the only record in the area they assigned this temperature to about 64,000 sq kilometers. I live in Stone Mountain Georgia where the average July temperatures are 88F (high) and 67F (low). Let's add 35 to that. That would mean that the average high would be 125 and the average low would be 102 for a whole month! The all time record high for Stone Mountain is 102 (according to weather.com, that sounds low to me). You don't question numbers like that? Why should you, GISS doesn't? And you call me an idiot. It's been 7 years and they still are showing this crap.
Now the USHCN is just US data. Maybe Haiti's data is better. How about that fine Soviet data or the Chinese, Indians, Brazilians or the Somalian data? Its bound to be accurate, right?
The reason I say the data was probably crap 50 years ago was that nobody really knows what it looked like 50 years ago. It was not designed to measure temperatures to within a tenth of a degree. These records were kept by postmasters, fireman, town clerks and newspapermen. To them within a couple of degrees was just fine. They would measure the temperature when they got around to it. It wasn't a job.
I'll tell you another bunch of old data I trust more then temperature. Sunspots. To measure temperature 100 years ago you need accurate measurements taken by thousands of individuals scattered around the planet all of which were willing to be as accurate as possible even though they weren't too worried about it and thats just on land. The oceans still wouldn't be covered. Who do you think took these temperatures? Sao Tome was a Portuguese colony in 1880 and Algeria was a French colony. Do you think that Sao Tome had a climatologist on staff? It was probably some rich plantation owner who did it for a hobby or the harbor master. At least we know it wasn't a slave. They had abolished slavery 4 years before! Most of the non-European and non-North American temperatures were taken in colonies. India has a few as does Australia and New Zealand. China zip. South America zip. Central America zip. Japan zip. Alaska 1 and Antarctica zip.
To measure sunspots 200 years ago you needed about a half a dozen astronomers around the world. One of them will have a clear sky every day and they kept detailed records. That doesn't mean the record is flawless just that it's better then the surface record. That's not much of a threshold to meet.