Signal to Noise Ratio in Measurement

Over at Climate Skeptic, I discuss Anthony Watt's preliminary findings as to the quality of measurement in the surface temperature installations that are used to measure global warming.  If we call global warming "the signal", then the signal is currently thought to have been about 0.6C over the last century.  However, Watt has good reason to estimate that 85% of the US Historical Climate Network has installation biases that create errors from 1-5C,or about 2-8 times the signal.  And these are not random biases that cancel out, but tend to all bias the numbers higher, leading to systematic over-estimation of temperature increases.