Professor Ben Herman Of The University Of Arizona on Maximum Temperature Trends

Here.

Excerpt:

There is an issue with regards to U.S. surface temperature trends that seems to have been overlooked, although apparently well recognized. I am referring to the HO-83 thermometers that were installed at many USHCN sites as well as first order stations. It has been well documented (Gall et.al. 1992, Jones et.al. 1995, Karl et. al. 1995) and others.that a warm bias existed, primarily in the daily maximum temperature readings reported by these instruments. The error in the Tucson data was about 2-3 deg F, but this error was apparently different with each thermometer. Karl et. al. (1995) have suggested that the average for this error over the country was on the order of 0.5 deg C on the reported maximum temperatures. Thus, if the maximum temperatures were corrected by this amount, average temperatures in the U.S, would be lowered by about 0.25 deg C, assuming the minimum readings were correct. This would probably pretty much neutralize the reported trend increase during the late 80’s and 90’s in this country.