“Accuracy of climate station electronic sensors – not the best“. Anthony Watts loves to talk about the accuracy of weather stations and temperature sensors. It’s an excellent way to distract from the fact that statistical analysis does a great job of detecting and removing the effect of such systemic issues.

This paper, Sensor and Electronic Biases/Errors in Air Temperature Measurements in Common Weather Station Networks by X. Lin and K. G. Hubbard gives him the perfect, and pointless, opportunity.

Look below at the chart Anthony posts as evidence and draw the obvious conclusion – over the range of common real-world conditions, the instrument error is flat! Even outside that range the worst-case inaccuracy is only on the order of 0.5%.

Notice the errors all happen away from common real-world temps.

It goes without saying (especially by Anthony) that instrumentation issues are closely watched and corrected for in the climate record, through the magic of statistical analysis.

P.S. Anthony, what is the difference between “accurate” and “consistent”? Which would be a better quality, for example, in a train station clock?