I live in a town in California where Firemen make $125K per year. They get 90% of their highest year's salary for retirement -- including overtime. Most of them work a ton of overtime in their last year and wind up with a pension of (on average) $150k per year. If a private citizen were to save up to to afford a 40 year retirement at $150K per year, they would need about $4M saved. So if it takes $4M to afford to retire like a fireman, then $7.5M seems about right to feel rich.

The Common Anomalies Method (CAM) requires that only those base stations that were present during the common base period be used when averaging the anomalies. Since this would be overly restrictive, the common practice at GISTemp is to apply manual estimated TOBS and MMTS adjustments as a way of artificially extending the series or infilling missing data. The method that GISTemp uses for making these adjustments is opaque and not adequately disclosed. But in all cases, such adjustments reintroduce (and perhaps amplify) errors that CAM was supposed to eliminate. Its Kabuki theater designed to create the appearance of science. Face it, there is no magically method for taking an ad hoc network of uncalibrated, poorly sited thermometers (whose location is constantly changing) of differing and changing technologies whose data was collected by random, untrained people at random and uncontrolled time of day and convert that into a global anomaly series capable of discerning temperature differences of 0.07C.

The errors in the individual measurements are not normally distributed because of UHI bias. And, Michael Mann, of all people, is now arguing that the temperature record is a *non-stationary* process (that's how he's now justifying his defective end-point padding)! So, if you accept that as true, then OLS regression is misspecified. Sorry, your grade school statistics won't work here. The error bars are immense.

No doubt its been getting warmer for the past 200 years. But, as an engineer, I question our ability to measure average global surface temperature to +/- 0.07 C. Such a measurement system would be an amazing engineering accomplishment. I don't question that the temperature anomaly for 2010 was 0.62, what I question is the asserted measurement error of +/- 0.07. I would accept a number of something like 0.62 +/- 0.50 with only a bit of skepticism. At +/- 0.25 my BS detector goes off. At +/-0.07 my rolling-on-the-floor-laughing-that-people-could-be-so-gullible reaction takes over.

If you can only observe 4.62% of the universe, how do you know that the other stuff, the stuff you can't see, isn't older?