It would be problematic for an agency established in 1970 to get "their temperature records" from 1918. Unless you mean another specific agency that existed at the time, but I don't think that's likely. If you want research from NOAA and other sources, it's been linked already.
OK, ask them for their research going back to 1970 because strangely enough, they leave out everything before 1981 to give the impression that Arctic ice is on a downward trend.
You're assuming that they immediately started measuring it after forming NOAA. Do you have a good reason for this? I've found no reason to believe they have their own measurements from before what's currently presented.
The International Panel on Climate Change (IPCC) has a 1990 report that uses NOAA data from 1974. It shows that Arctic sea ice is cyclical, as it's always been, and had a peak in 1979. SO NOAA was measuring sea ice extent at least as far back as 1974.
Also, the third graph on page 202 shows that the Earth was MUCH warmer a thousand years ago than it is now, despite far less CO2. This isn't particularly surprising since carbon dioxide levels track global temperature - not the other way around - as the ice core data going back 800 000 years shows.
The 1985 Department of Energy report Projecting the Climatic Effects of Increasing Carbon Dioxide uses data going back to 1925 that shows, surprise, that Arctic sea ice is normal, cyclical and has nothing to do with CO2 levels.
> Also, the third graph on page 202 shows that the Earth was MUCH warmer a thousand years ago than it is now, despite far less CO2.
It's true, but doesn't prove your point. Specifically, we're still left with two options: what we see now is the same cyclic effect, or what we see now is caused by us and if we get another cycle on top we're screwed even faster. You'd have to show the inverse - that the production of lots of CO2 does not affect the cyclic process. (Or that the pattern of the current raise is extremely likely to belong to a cycle)